MF研究者総覧

Researcher Database

Advance
Masaharu YOSHIOKA
Faculty of Information Science and Technology　Computer Science and Information Technology　Knowledge Software Science
Professor
Researchers' Page on researchmap
Last Updated :2024/05/09

Researcher Profile and Settings

Affiliation

  • Faculty of Information Science and Technology　Computer Science and Information Technology　Knowledge Software Science

Job Title

  • Professor

URL

J-Global ID

Research Interests

  • Nanoinformatics   テキストマイニング   検索者の意図   評価実験   マルチファセット   地理情報   意見分析   設計文書   相関性分析   形式的知識   インタフェース   検索システムの評価   文書管理   フォーマルコンセプト   設計対象モデル   クラスタリング   XML   問題解決   知識の再利用   検索拡張   知識メディア   極大類比   検索要求   アブダクション   セマンティックWeb   アナジロジー   文書構造   設計学   アナロジー   思考過程モデル   国際標準化   検索語拡張   概念の汎化   ストーリー   設計知識   知識マネージメント   オントロジー   検索意図   適合的汎化   シソーラス   電子化辞書   情報検索   

Research Areas

  • Informatics / Human interfaces and interactions
  • Informatics / Database science
  • Informatics / Intelligent informatics

Academic & Professional Experience

  • 2019/04 - Today Hokkaido University Faculty of Information Science and Technology Division of Computer Science and Information Technology
  • 2018/01 - 2019/03 Hokkaido University Graduate School of Information Science and Technology Division of Computer Science and Information Technology Professor
  • 2014/04 - 2017/12 Hokkaido University Graduate School of Information Science and Technology, Division of Computer Science and Information Technology Associate Professor
  • 2008/04 - 2014/03 Hokkaido University Graduate School of Information Science and Technology Division of Computer Science
  • 2004/04 - 2008/03 Hokkaido University Graduate School of Information Science and Technology Division of Computer Science Associate Professor
  • 2001/06 - 2004/03 Hokkaido University Graduate School of Engineering
  • 2000/04 - 2001/05 National Institute of Informatics
  • 1996/04 - 2000/03 The National Center for Science Information Systems Research Associate

Association Memberships

  • ACM (Association for Computing Machinery)   言語処理学会   日本機械学会   精密工学会   人工知能学会   情報処理学会   

Research Activities

Published Papers

  • Kojiro Machi, Seiji Akiyama, Yuuya Nagata, Masaharu Yoshioka
    Journal of chemical information and modeling 63 (21) 6619 - 6628 2023/10/19 [Refereed]
     
    There is a pressing need for the automated extraction of chemical reaction information because of the rapid growth of scientific documents. The previously reported works in the literature for the procedure extraction either (a) did not consider the semantic relations between the action and argument or (b) defined a detailed schema for the extraction. The former method was insufficient for reproducing the reaction, while the latter methods were too specific to their own schema and did not consider the general semantic relation between the verb and argument. In addition, they did not provide an annotated text that aligned with the structured procedure. Along these lines, in this work, we propose a corpus named organic synthesis procedures with argument roles (OSPAR) that is annotated with rolesets to consider the semantic relation between the verb and argument. We also provide rolesets for chemical reactions, especially for organic synthesis, which represent the argument roles of actions in the corpus. More specifically, we annotated 112 organic synthesis procedures in journal articles from Organic Syntheses and defined 19 new rolesets in addition to 29 rolesets from an existing language resource (Proposition Bank). After that, we constructed a simple deep learning system trained on OSPAR and discussed the usefulness of the corpus by comparing it with chemical description language (XDL) generated by a natural language processing tool, namely, SynthReader. While our system's output required more detailed parsing, it covered comparable information against XDL. Moreover, we confirmed that the validation of the output action sequence was easy as it was aligned with the original text.
  • Mi-Young Kim, Juliano Rabelo, Randy Goebel, Masaharu Yoshioka, Yoshinobu Kano, Ken Satoh
    New Frontiers in Artificial Intelligence 51 - 67 0302-9743 2023/04/06
  • Randy Goebel, Yoshinobu Kano, Mi-Young Kim, Juliano Rabelo 0001, Ken Satoh, Masaharu Yoshioka
    ICAIL 472 - 480 2023
  • Yasuhiro Aoki, Masaharu Yoshioka, Youta Suzuki
    REVIEW OF SOCIONETWORK STRATEGIES 2523-3173 2022/02 [Refereed]
     
    A legal textual entailment task is a task to recognize entailment between a law article and its statements. In the Competition on Legal Information Extraction/Entailment (COLIEE), this task is designed as a task to confirm the entailment of a yes/no answer from the given civil code article(s). Based on the development of deep-learning-based natural language processing tools such as bidirectional encoder representations from transformers (BERT), many participants in the task used such tools, and the best performance system of COLIEE 2020 was a BERT-based system. However, because of the limitation of the size of training data provided by the task organizer, training such tools to adapt to the variability of the questions is difficult. In this paper, we propose a data-augmentation method to make training data using civil code articles for understanding the syntactic structure of the questions and articles for entailment. Our BERT-based ensemble system, which uses this augmentation method, achieves the best performance (accuracy = 0.7037) in Task 4 of COLIEE 2021. We also introduce the results of additional experiments to discuss the characteristics of the proposed method.
  • Juliano Rabelo, Randy Goebel, Mi-Young Kim, Yoshinobu Kano, Masaharu Yoshioka, Ken Satoh
    REVIEW OF SOCIONETWORK STRATEGIES 2523-3173 2022/02 [Refereed]
     
    We summarize the 8th Competition on Legal Information Extraction and Entailment. In this edition, the competition included five tasks on case law and statute law. The case law component includes an information retrieval Task (Task 1), and the confirmation of an entailment relation between an existing case and an unseen case (Task 2). The statute law component includes an information retrieval Task (Task 3), an entailment/question answering task based on retrieved civil code statutes (Task 4) and an entailment/question answering task without retrieved civil code statutes (Task 5). Participation was open to any group based on any approach. Eight different teams participated in the case law competition tasks, most of them in more than one task. We received results from six teams for Task 1 (16 runs) and 6 teams for Task 2 (17 runs). On the statute law task, there were eight different teams participating, most in more than one task. Six teams submitted a total of 18 runs for Task 3, 6 teams submitted a total of 18 runs for Task 4, and 4 teams submitted a total of 12 runs for Task 5. Here we summarize the approaches, our official evaluation, and analysis on our data and submission results.
  • Kojiro Machi, Masaharu Yoshioka
    CLEF (Working Notes) 797 - 807 2022
  • Yuan Li, Biaoyan Fang, Jiayuan He 0002, Hiyori Yoshikawa, Saber A. Akhondi, Christian Druckenbrodt, Camilo Thorne, Zubair Afzal, Zenan Zhai, Kojiro Machi, Masaharu Yoshioka, Youngrok Jang, Hosung Song, Junho Lee, Gyeonghun Kim, Yireun Kim, Stanley Jungkyu Choi, Honglak Lee, Kyunghoon Bae, Darshini Mahendran, Christina Tang, Bridget T. McInnes, Timothy Baldwin, Karin Verspoor
    CLEF (Working Notes) 758 - 781 2022
  • Interactive Visualization of;the Polarity-based;Stance of News;Websites using News;Genres
    Masaharu Yoshioka, Norihiko Tatsunami, Masahiko Itoh, Noriko Kando, James Allan
    Proceedings of the 9th International Workshop on News Recommendation and Analytics (INRA 2021) 2021/09 [Refereed][Not invited]
  • Motohiro Sunouchi, Masaharu Yoshioka
    IVSP 2021: 3rd International Conference on Image, Video and Signal Processing(IVSP) 65 - 72 2021/03 [Refereed]
  • Juliano Rabelo, Mi Young Kim, Randy Goebel, Masaharu Yoshioka, Yoshinobu Kano, Ken Satoh
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12758 LNAI 196 - 210 0302-9743 2021 [Refereed]
     
    We present a summary of the 7th Competition on Legal Information Extraction and Entailment. The competition consists of four tasks on case law and statute law. The case law component includes an information retrieval task (Task 1), and the confirmation of an entailment relation between an existing case and an unseen case (Task 2). The statute law component includes an information retrieval task (Task 3) and an entailment/question answering task (Task 4). Participation was open to any group based on any approach. Ten different teams participated in the case law competition tasks, most of them in more than one task. We received results from 9 teams for Task 1 (22 runs) and 8 teams for Task 2 (22 runs). On the statute law task, there were 14 different teams participating, most in more than one task. Eleven teams submitted a total of 28 runs for Task 3, and 13 teams submitted a total of 30 runs for Task 4. We summarize the approaches, our official evaluation, and analysis on our data and submission results.
  • Motohiro SUNOUCHI, Masaharu YOSHIOKA
    IEICE Transactions on Information and Systems E104D (10) 1734 - 1748 0916-8532 2021 [Refereed]
     
    This paper proposes new acoustic feature signatures based on the multiscale fractal dimension (MFD), which are robust against the diversity of environmental sounds, for the content-based similarity search. The diversity of sound sources and acoustic compositions is a typical feature of environmental sounds. Several acoustic features have been proposed for environmental sounds. Among them is the widely-used Mel-Frequency Cepstral Coefficients (MFCCs), which describes frequency-domain features. However, in addition to these features in the frequency domain, environmental sounds have other important features in the time domain with various time scales. In our previous paper, we proposed enhanced multiscale fractal dimension signature (EMFD) for environmental sounds. This paper extends EMFD by using the kernel density estimation method, which results in better performance of the similarity search tasks. Furthermore, it newly proposes another acoustic feature signature based on MFD, namely very-long-range multiscale fractal dimension signature (MFD-VL). The MFD-VL signature describes several features of the time-varying envelope for long periods of time. The MFD-VL signature has stability and robustness against background noise and small fluctuations in the parameters of sound sources, which are produced in field recordings. We discuss the effectiveness of these signatures in the similarity sound search by comparing with acoustic features proposed in the DCASE 2018 challenges. Due to the unique descriptiveness of our proposed signatures, we confirmed the signatures are effective when they are used with other acoustic features.
  • Masaharu Yoshioka, Yasuhiro Aoki, Youta Suzuki
    Proceedings of the 18th International Conference on Artificial Intelligence and Law, ICAIL 2021 278 - 284 2021 [Refereed]
     
    The Competition on Legal Information Extraction/Entailment (COLIEE) statute law legal textual entailment task (task 4) is a task to make a system judge whether a given question statement is true or not by provided articles. In the last COLIEE 2020, the best performance system used bidirectional encoder representations from transformers (BERT), a deep-learning-based natural language processing tool for handling word semantics by considering their context. However, there are problems related to the small amount of training data and the variability of the questions. In this paper, we propose a BERT-based ensemble method with data augmentation to solve this problem. For the data augmentation, we propose a systematic method to make training data for understanding the syntactic structure of the questions and articles for entailment. In addition, due to the nature of the non-deterministic characteristics of BERT fine-tuning and the variability of the questions, we propose a method to construct multiple BERT fine-tuning models and select an appropriate set of models for ensemble. The accuracy of our proposed method for task 4 was 0.7037, which was the best performance among all submissions.
  • Kojiro Machi, Masaharu Yoshioka
    CEUR Workshop Proceedings 2936 720 - 731 1613-0073 2021 [Not refereed]
     
    This paper describes our system for the ChEMU 2021 task 2 of anaphora resolution for extracting chemical reactions from patents. We divide the task into two subtasks: (1) span detection of mentions that include antecedent and anaphor; and (2) mention classification and relation detection. For the first task, we use a deep learning-based approach. For the second task, we use a rule-based approach that uses features related to chemical reactions by ChemicalTagger, which is a state-of-the-art text-mining tool for chemistry. Our system obtained an exact-match F-score of 0.6907 and a relaxed match F-score of 0.7459.
  • Masaharu Yoshioka, Hideo Joho
    Information Retrieval Series 43 127 - 141 1871-7500 2021 [Invited]
     
    This chapter introduces the research background and details of temporal information access tasks in the NTCIR. The GeoTime task was the first attempt to evaluate temporal information retrieval as an extension of an information-retrieval-for-question-answering task. Temporalia was a task to investigate the role of temporal factors in a search.
  • Masaharu Yoshioka, Takanori Nakagawa
    CEUR Workshop Proceedings 2721 300 - 305 1613-0073 2020 [Refereed]
     
    Wikipedia categories are intended to group together pages on similar topics and are organized in a hierarchical structure. Since the editorial policy for Wikipedia categories differs from policies used by knowledge engineers, various types of relationship exist in its category structure. In this paper, we propose a novel framework called "Wikipedia category ontology" (WCO) that aims to act as a basis for interpreting the Wikipedia category structure and is based on a classification of category and relationship types. WCO enables particular Wikipedia category substructures to be extracted, including class-subclass hierarchies, class-instance, and sets of diffused categories for a particular category. WCO is available online in the form of linked open data at http://wcontology.org/.
  • 狩野芳伸, 吉岡真治, RABELO Juliano, KIM Mi-Young, GOEBEL Randy, 佐藤健
    Journal of the Japanese Society for Artificial Intelligence 一般社団法人 人工知能学会 35 (3) 377 - 384 2188-2266 2020 [Not refereed][Invited]
  • Juliano Rabelo, Mi-Young Kim, Randy Goebel, Masaharu Yoshioka, Yoshinobu Kano, Ken Satoh
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12331 LNAI 34 - 49 0302-9743 2019 [Refereed][Not invited]
     
    We summarize the evaluation of the 6th Competition on Legal Information Extraction/Entailment (COLIEE 2019). The competition consists of four tasks: two on case law and two on statute law. The case law component includes an information retrieval task (Task 1), and the confirmation of an entailment relation between an existing case and an unseen case (Task 2). The statute law component also includes an information retrieval task (Task 3) and an entailment/question answering task (Task 4), which attempts to confirm whether a particular statute applies to a yes/no question. Participation was open to any group in the world, based on any approach. Eleven different teams participated in the case law competition tasks, some of them in more than one task. We received results from 7 teams for Task 1 (15 runs) and 7 teams for Task 2 (18 runs). For the statute law tasks, 8 different teams participated, some in more than one task. Seven teams submitted a total of 13 runs for Task 3, and 7 teams submitted a total of 15 runs for Task 4. Here we summarize each team’s approaches, our official evaluation, and some analysis of the variety of methods that produced the evaluation results.
  • Construction of In-House Papers/Figures Database System using PDFs
    M. Yoshioka, S. Hara
    Collected Abstract of the 12th International Workshop on Information Search, Integration, and Personalization (ISIP 2018), Fukuoka, Japan, May 14-15, 2018. 1 - 1 2018/05 [Refereed][Not invited]
  • Yoshinobu Kano, Mi-Young Kim, Masaharu Yoshioka, Yao Lu, Juliano Rabelo, Naoki Kiyota, Randy Goebel, Ken Satoh
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 11717 LNAI 177 - 192 0302-9743 2018 [Refereed][Not invited]
     
    We summarize the evaluation of the 5th Competition on Legal Information Extraction/Entailment 2018 (COLIEE-2018). The COLIEE-2018 tasks include two tasks in each of statute law and case law. The case law component includes an information retrieval (Task 1), and the confirmation of an entailment relation between an existing case and an unseen case (Task 2). The statute law component includes information retrieval (Task 3) and entailment/question answering (Task 4). Participation was open to any group based on any approach. 13 teams participated in the case law competition, and we received results from 7 teams where 6 submissions to Task 1 (12 runs), and 4 submissions to Task 2 (8 runs). Regarding the statute law, there were submissions of 17 runs from 8 teams (including 2 organizers’ runs) for Task 3 and 7 runs from 3 teams for Task 4. We describe each team’s approaches, our official evaluation, and analysis on our data and submission results. We also discuss possibilities for future competition tasks.
  • Masaharu Yoshioka, Shinjiro Hara
    Information Search, Integration, and Personalization (Communications in Computer and Information Science (CCIS), edited by D. Kotzinos, D. Laurent, N. Spyratos, Y. Tanaka, and R. Taniguchi, Springer International Publishing, Switzerland, 2019) Vol. 1040 41 - 52 1865-0929 2018 [Refereed][Not invited]
     
    Several general-purpose databases of research papers are available, including ScienceDirect and Google Scholar. However, these systems may not include recent papers from workshops or conferences for special interest groups. Moreover, it can be helpful for researchers in a particular domain to analyze research papers using their own terminology. To support these researchers, we propose a new database system based on the information extracted from portable document format documents that enables annotation of terms via a terminology extraction system. In this paper, we evaluate our system using a use-case experiment and discuss the appropriateness of the system.
  • Masaharu Yoshioka, Myungha Jang, James Allan, Noriko Kando
    Proceedings of the Second International Workshop on Recent Trends in News Information Retrieval co-located with 40th European Conference on Information Retrieval (ECIR 2018), Grenoble, France, March 26, 2018. CEUR-WS.org 2079 6 - 8 1613-0073 2018 [Refereed][Not invited]
     
    We develop a novel framework that helps identify potential bias in news websites to support users who are exposed to news articles with a wide variety of political leanings. We propose a polarity-based stance (PS), a vector that represents how often a website publishes articles that are positive or negative with regard to a topic. We derive PS using the GDELT database and visualize the news websites' stances. We demonstrate the utility of our framework via a case study of the 2016 US Presidential Election.
  • Masaharu Yoshioka
    NEW FRONTIERS IN ARTIFICIAL INTELLIGENCE (JSAI-ISAI 2017) 10838 5 - 19 0302-9743 2018 [Refereed][Not invited]
     
    The Competition on Legal Information Extraction/Entailment (COLIEE) involves the legal question answering task. The information retrieval task for finding relevant articles to questions is one of the subtasks in COLIEE. In this paper, we compare the characteristics of the test data provided in two different language (English and Japanese) and analyze topic difficulty based on the submission data by using the retrieval results of Indri, a state-of-the-art information retrieval system. We also discuss issues relating to the design of new COLIEE information retrieval tasks.
  • Hayato Shiokawa, Kota Kawaguchi, Bingcai Han, Takehito Utsuro, Yasuhide Kawada, Masaharu Yoshioka, Noriko Kando
    NATURAL LANGUAGE PROCESSING TECHNIQUES FOR EDUCATIONAL APPLICATIONS 143 - 151 2018 [Refereed][Not invited]
     
    Search engine is an important tool of modern academic study, but the results are lack of measurement of beginner friendliness. In order to improve the efficiency of using search engine for academic study, it is necessary to invent a technique of measuring the beginner friendliness of a Web page explaining academic concepts and to build an automatic measurement system. This paper studies how to integrate heterogeneous features such as a neural image feature generated from the image of the Web page by a variant of CNN (convolutional neural network) as well as text features extracted from the body text of the HTML file of the Web page. Integration is performed through the framework of the SVM classifier learning. Evaluation results show that heterogeneous features perform better than each individual type of features.
  • Masaharu Yoshioka
    Workshop and Poster Proceedings of the 8th Joint International Semantic Technology Conference co-located with the 8th Joint International Semantic Technology Conference (JIST 2018), Awaji City, Hyogo, Japan, November 26-28, 2018. CEUR-WS.org 2293 98 - 101 1613-0073 2018 [Refereed][Not invited]
     
    We demonstrate Wikipedia Category Comprehensiveness Checker (WC3) based on the DBpedia metadata database. This system supports to check comprehensiveness and consistency of Wikipedia category annotation using DBpedia database. This system is available online.
  • Construction of In-House Papers/Figures Database System for a Paticular Research Domain using PDFs - Application of Nano-Crystal Device Development Domain -
    M. Yoshioka, S. Hara
    Proccedings of the 2nd International Workshop on Scientific Document Analysis (SCIDOCA 2017), Tokyo, Japan, November 14-15, 2017: Proccedings of the 2nd International Workshop on Scientific Document Analysis (7pp.) 1 - 7 2017/11 [Refereed][Not invited]
  • Masaharu Yoshioka
    KNOWLEDGE GRAPHS AND LANGUAGE TECHNOLOGY 10579 119 - 136 0302-9743 2017 [Refereed][Not invited]
     
    WC3 (Wikipedia Category Consistency Checker) is a system that supports the analysis of the metadata-annotation style in Wikipedia articles belonging to a particular Wikipedia category (the subcategory of "Categories by parameter") by using the DBpedia metadata database. This system aims to construct an appropriate SPARQL query to represent the category and compares the retrieved results and articles that belong to the category. In this paper, we introduce WC3 and extend the algorithm to analyze efficiently additional varieties of Wikipedia category. We also discuss the metadata-annotation quality of the Wikipedia by usingWC3. URL ofWC3 is http://wnews.ist.hokudai.ac.jp/wc3/ and related files are available at http://wnews.ist.hokudai.ac.jp/wc3/KEKI.
  • Masaharu Yoshioka, Daiki Onodera
    COLIEE 2017. 4th Competition on Legal Information Extraction and Entailment, held in conjunction with the 16th International Conference on Artificial Intelligence and Law (ICAIL 2017) in King's College London, UK. 9 - 22 2017 [Refereed][Not invited]
  • A multi-faceted figure retrieval system from research papers for supporting nano-crystal device development researchers
    Masaharu Yoshioka, Tao Zhu, Shinjiroh Hara
    The Proceedings of the First International Workshop on ScIentific Document Analysis (SCIDOCA 2016) 2016/11 [Refereed][Not invited]
  • Civil Code Article Information Retrieval System based on Legal Terminology and Civil Code Article Structure
    Daiki Onodera, Masaharu Yoshioka
    The Proceedings of the 10th International Workshop on Juris-Informatics (JURISIN2016), The Japanese Society of Artificial Intelligence 2016/11 [Refereed][Not invited]
  • WC3: Analyzing the style of metadata annotation among Wikipedia articles by using Wikipedia category and the DBPedia metadata database.
    Masaharu YOSHIOKA
    Proceedings of the First Workshop on Knowledge Extraction and Knowledge Integration 2016/10 [Refereed][Not invited]
  • Masaharu Yoshioka, Masahiko Itoh, Michele Sebag
    INFORMATION SEARCH, INTEGRATION, AND PERSONALIZATION, (ISIP 2015) 622 142 - 156 1865-0929 2016 [Refereed][Not invited]
     
    Data visualization is a core approach for understanding data specifics and extracting useful information in a simple and intuitive way. Visual data mining proceeds by projecting multidimensional data onto two-dimensional (2D) or three-dimensional (3D) data, e.g., through mathematical optimization and topology preserved in multidimensional scaling (MDS). However, this projection does not necessarily comply with the user's needs, prior knowledge and/or expectations. This paper proposes an interactive visual mining approach, centered on the user's needs and allowing the modification of data visualization by leveraging approaches from metric learning. The paper exemplifies the proposed system, referred to as Interactive Metric Learning-based Visual Data Exploration (IMViDE), applied to scientific social network browsing.
  • Masaharu Yoshioka, Noriko Kando
    Proceedings of the First International Workshop on Recent Trends in News Information Retrieval co-located with 38th European Conference on Information Retrieval (ECIR 2016), Padua, Italy, March 20, 2016. CEUR-WS.org 1568 63 - 65 1613-0073 2016 [Refereed][Not invited]
     
    The News Site Contrast (NSContrast) system analyzes news articles retrieved from multiple news sites based on the concept of contrast set mining. It can extract terms that characterize different topics of interest across news sites, countries, and regions. In this study, we used NSContrast to analyze Global Database of Events, Language, and Tone (GDELT) data by comparing news articles from different regions (e.g., USA, Asia, and the Middle East). We also present examples of analyses performed using this system.
  • Thaer M. Dieb, Masaharu Yoshioka, Shinjiro Hara
    Journal of Information Processing 24 (3) 554 - 564 1882-6652 2016 [Refereed][Not invited]
     
    The process of nanocrystal device development is not well systematized. To support this process, analysis of the information produced by developmental experiments is required. In this study, we constructed an annotated corpus to support the extraction of experimental information from relevant publications. We designed the corpus-construction guidelines by cooperating with a domain expert. We evaluated these guidelines through corpusconstruction experiments with graduate students from this domain, and then evaluated the corpus with the domain expert. In the corpus construction experiments, we achieved a sufficient level of Inter-Annotator Agreement by using a loose agreement measure that ignored the term-boundary mismatch problem, and made an agreement corpus that excluded annotations based on misunderstanding the guidelines. The domain expert evaluated this agreement corpus and modified the guidelines based on real examples. Using these guidelines, we finalized the corpus called "NaDev" (Nanocrystal Device development corpus). The NaDev corpus and its construction guidelines will be released via our website, http://nanoinfo.ist.hokudai.ac.jp/. The NaDev corpus aims to support automatic information extraction from publications relevant to nanocrystal device development. This information can be used to solve problems in the nanotechnology domain using the massive availability of fresh information. To the best of our knowledge, this is the first corpus constructed for the development of nanocrystal devices.
  • Thaer M. Dieb, Masaharu Yoshioka
    Transactions on Machine Learning and Data Mining 8 (2) 61 - 76 2015/10 [Refereed][Not invited]
  • Thaer M. Dieb, Masaharu Yoshioka, Shinjiro Hara, Marcus C. Newton
    BEILSTEIN JOURNAL OF NANOTECHNOLOGY 6 1872 - 1882 2190-4286 2015/09 [Refereed][Not invited]
     
    To support nanocrystal device development, we have been working on a computational framework to utilize information in research papers on nanocrystal devices. We developed an annotated corpus called "NaDev" (Nanocrystal Device Development) for this purpose. We also proposed an automatic information extraction system called "NaDevEx" (Nanocrystal Device Automatic Information Extraction Framework). NaDevEx aims at extracting information from research papers on nanocrystal devices using the NaDev corpus and machine-learning techniques. However, the characteristics of NaDevEx were not examined in detail. In this paper, we conduct system evaluation experiments for NaDevEx using the NaDev corpus. We discuss three main issues: system performance, compared with human annotators; the effect of paper type (synthesis or characterization) on system performance; and the effects of domain knowledge features (e.g., a chemical named entity recognition system and list of names of physical quantities) on system performance. We found that overall system performance was 89% in precision and 69% in recall. If we consider identification of terms that intersect with correct terms for the same information category as the correct identification, i.e., loose agreement (in many cases, we can find that appropriate head nouns such as temperature or pressure loosely match between two terms), the overall performance is 95% in precision and 74% in recall. The system performance is almost comparable with results of human annotators for information categories with rich domain knowledge information (source material). However, for other information categories, given the relatively large number of terms that exist only in one paper, recall of individual information categories is not high (39-73%); however, precision is better (75-97%). The average performance for synthesis papers is better than that for characterization papers because of the lack of training examples for characterization papers. Based on these results, we discuss future research plans for improving the performance of the system.
  • Takehito Utsuro, Yusuke Inoue, Takakazu Imada, Masaharu Yoshioka, Noriko Kando
    Social Media for Government Services 135 - 149 2015/01/01 [Refereed][Not invited]
     
    This chapter presents a framework of detecting bursty topics of correlated news and twitter, and discusses how to integrate the framework into government services. Especially, as a specific application of the proposed framework of detecting bursty topics of correlated news and twitter, this chapter gives an example of collecting news and twitter that are related to “the 2012 London Olympic game” and applying the proposed framework.
  • Masaharu Yoshioka, Rhett Loban
    2015 11TH INTERNATIONAL CONFERENCE ON SIGNAL-IMAGE TECHNOLOGY & INTERNET-BASED SYSTEMS (SITIS) 712 - 718 2015 [Refereed][Not invited]
     
    Wikipedia is a free Internet encyclopedia that covers numerous topics and is maintained by voluntary editors. There have been several efforts to analyze the quality of Wikipedia's content, specifically assessing textual content and user editorial behavior. However, there has been no specific analysis of the quality of structured data (e.g., category structures, attributes contained in infoboxes) in Wikipedia. DBPedia is one of a core part of Linked Open Data database which extracts structured information from Wikipedia. Since the quality of DBPedia depends on Wikipedia, it is beneficial to have a mechanism to keep a consistent style of writing of the structured data (metadata) within Wikipedia. In this paper, we propose a system WC3 (WC-triple: Wikipedia Category Consistency Checker) that uses DBPedia RDF-triple information to evaluate the consistency of the metadata related to the Wikipedia category information. We also illustrate examples of typical inconsistencies that the system can assist in detecting.
  • Martin Krallinger, Obdulia Rabal, Florian Leitner, Miguel Vazquez, David Salgado, Zhiyong Lu, Robert Leaman, Yanan Lu, Donghong Ji, Daniel M. Lowe, Roger A. Sayle, Riza Theresa Batista-Navarro, Rafal Rak, Torsten Huber, Tim Rocktaschel, Serergio Matos, David Campos, Buzhou Tang, Hua Xu, Tsendsuren Munkhdalai, Keun Ho Ryu, S. V. Ramanan, Senthil Nathan, Slavko Zitnik, Marko Bajec, Lutz Weber, Matthias Irmer, Saber A. Akhondi, Jan A. Kors, Shuo Xu, Xin An, Utpal Kumar Sikdar, Asif Ekbal, Masaharu Yoshioka, Thaer M. Dieb, Miji Choi, Karin Verspoor, Madian Khabsa, C. Lee Giles, Hongfang Liu, Komandur Elayavilli Ravikumar, Andre Lamurias, Francisco M. Couto, Hong-Jie Dai, Richard Tzong-Han Tsai, Caglar Ata, Tolga Can, Anabel Usie, Rui Alves, Isabel Segura-Bedmar, Paloma Martinez, Julen Oyarzabal, Alfonso Valencia
    JOURNAL OF CHEMINFORMATICS 7 (S-1) S2  1758-2946 2015/01 [Refereed][Not invited]
     
    The automatic extraction of chemical information from text requires the recognition of chemical entity mentions as one of its key steps. When developing supervised named entity recognition (NER) systems, the availability of a large, manually annotated text corpus is desirable. Furthermore, large corpora permit the robust evaluation and comparison of different approaches that detect chemicals in documents. We present the CHEMDNER corpus, a collection of 10,000 PubMed abstracts that contain a total of 84,355 chemical entity mentions labeled manually by expert chemistry literature curators, following annotation guidelines specifically defined for this task. The abstracts of the CHEMDNER corpus were selected to be representative for all major chemical disciplines. Each of the chemical entity mentions was manually labeled according to its structure-associated chemical entity mention (SACEM) class: abbreviation, family, formula, identifier, multiple, systematic and trivial. The difficulty and consistency of tagging chemicals in text was measured using an agreement study between annotators, obtaining a percentage agreement of 91. For a subset of the CHEMDNER corpus (the test set of 3,000 abstracts) we provide not only the Gold Standard manual annotations, but also mentions automatically detected by the 26 teams that participated in the BioCreative IV CHEMDNER chemical mention recognition task. In addition, we release the CHEMDNER silver standard corpus of automatically extracted mentions from 17,000 randomly selected PubMed abstracts. A version of the CHEMDNER corpus in the BioC format has been generated as well. We propose a standard for required minimum information about entity annotations for the construction of domain specific corpora on chemical and drug entities. The CHEMDNER corpus and annotation guidelines are available at: http://www.biocreative.org/resources/biocreative-iv/chemdner-corpus/
  • Automatic Annotation of Parameters from Nanodevice Development Research Papers
    T. M. Dieb, M. Yoshioka, S. Hara, M. C. Newton
    Proceedings of the 4th International Workshop on Computational Terminology (COMPUTERM 2014), COLING 2014 Workshop, Dublin, Ireland, August 23, 2014: Proceedings of the 4th International Workshop on Computational Terminology, pp. 77-85 (9pp.) 77 - 85 2014/08 [Refereed][Not invited]
  • Masaharu Yoshioka
    INFORMATION RETRIEVAL TECHNOLOGY, AIRS 2014 8870 470 - 481 0302-9743 2014 [Refereed][Not invited]
     
    Wikipedia is a free encyclopedia on the Internet that contains millions of articles. Wikipedia ontologies (e.g., YAGO2 and the Japanese Wikipedia ontology) have been constructed to utilize the semantic information in this encyclopedia; in these ontologies, the category information from each article is used to construct concept hierarchies. However, Wikipedia categories were originally designed for navigation, to help users find appropriate articles, and they are not equivalent to concept hierarchies in Wikipedia ontologies. In this study, we briefly review the definition of a category in Wikipedia and exhaustively investigate the category structure in the Japanese Wikipedia. We also discuss how category structure information can be utilized to construct an ontology and to calculate a semantic similarity measure.
  • Ensemble Approach to Extract Chemical Named Entity
    Masaharu Yoshioka, Thaer M. Dieb
    Proceedings of the fourth BioCreative challenge evaluation workshop, vol. 2, 162 - 167 2013/10 [Not refereed][Not invited]
  • Bursty Topics in Time Series Japanese / Chinese News Streams and their Cross-Lingual Alignment
    Liyi Zheng, Takehito Utsuro, Masaharu Yoshioka
    Proceedings of the 13th Conference of the Pacific Association for Computational Linguistics 2013/09 [Refereed][Not invited]
  • Masaharu YOSHIOKA
    Proceedings of the 10th NTCIR Workshop Meeting on Evaluation of Information Access Technologies: Information Retrieval, Quesiton Answering, And Cross-Lingual Information Access, National Institute of Informatics (NII) 255 - 259 2013/06 [Refereed][Not invited]
  • Masaharu Yoshioka, Noriko Kando
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 7774 375 - 381 0302-9743 2013 [Refereed][Not invited]
     
    It is now possible to use various geographical open data sources such as GeoNames and Wikipedia to construct geographic information systems. In addition, these open data sources are integrated by the concept of Linked Open Data. There have been several attempts to identify links between existing data, but few studies have focused on the quality of such links. In this paper, we introduce an automatic link discovery method for identifying the correspondences between GeoNames entries and Wikipedia pages, based on Wikipedia category information. This method finds not only appropriate links but also inconsistencies between two databases. Based on this integration results, we discuss the type of inconsistencies for making consistent Linked Open Data. © Springer-Verlag 2013.
  • Daichi Koike, Yusuke Takahashi, Takehito Utsuro, Masaharu Yoshioka, Noriko Kando
    Sixth International Joint Conference on Natural Language Processing, IJCNLP 2013, Nagoya, Japan, October 14-18, 2013 Asian Federation of Natural Language Processing / ACL 917 - 921 2013 [Refereed][Not invited]
  • Masaharu Yoshioka, Takahiro Fujiwara
    Proceedings of the 7th Workshop on Geographic Information Retrieval, GIR 2013 57 - 63 2013 [Refereed][Not invited]
     
    When processing toponym information in natural language text, it is crucial to have a good gazetteers. There are several well-organized gazetteers for English text, but they do not cover Japanese local toponyms. In this paper, we introduce a Japanese gazetteers based on Open Data (e.g., the Toponym database distributed by Japanese ministries, Wikipedia, and GeoNames) and propose a toponym disambiguation framework that uses the constructed gazetteers. We also evaluate our approach based on a blog corpus that contains place names with high ambiguity.
  • Issues for Linking Geographical Open Data of GeoNames and Wikipedia
    Masaharu Yoshioka, Noriko Kando
    Proceedings of the 2nd Joint International Semantic Technology Conference (JIST2012), 2012/12 [Refereed][Not invited]
  • Cross-Lingual Topic Alignment in Time Series Japanese / Chinese News
    Shuo Hu, Yusuke Takahashi, Liyi Zheng, Takehito Utsuro, Masaharu Yoshioka, Noriko Kando, Tomohiro Fukuhara, Hiroshi Nakagawa, Yoji Kiyota
    Proceedings of the 26th Pacific Asia Conference on Language, Information and Computation 532 - 541 2012/11 [Refereed][Not invited]
  • Masaharu Yoshioka, Yiqi Liu, Noriko Kando
    情報処理学会論文誌. データベース 一般社団法人情報処理学会 5 (3) 141 - 148 0387-5806 2012/09 [Refereed][Not invited]
     
    Recently, due to the higher demand for geographic information system, it is necessary to have a good geographical database for such systems. GeoNames is one of the largest geographical database as Open Data. This database is also used for constructing web ontology by adding links to the Wikipedia page as a part of Linked Open Data. Even though, here are several attempts to find links automatically, the number of links between GeoNames and Wikipedia is not sufficient. In this paper, we propose an automatic link discovery method to use Wikipedia categories to identify the correspondence betwe...
  • Yusuke Takahashi, Takehito Utsuro, Masaharu Yoshioka, Noriko Kando, Tomohiro Fukuhara, Hiroshi Nakagawa, Yoji Kiyota
    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 7614 239 - 249 0302-9743 2012 [Refereed][Not invited]
     
    This paper focuses on two types of modeling of information flow in news stream, namely, burst analysis and topic modeling. First, when one wants to detect a kind of topics that are paid much more attention than usual, it is usually necessary for him/her to carefully watch every article in news stream at every moment. In such a situation, it is well known in the field of time series analysis that Kleinberg's modeling of bursts is quite effective in detecting burst of keywords. Second, topic models such as LDA (latent Dirichlet allocation) are also quite effective in estimating distribution of topics over a document collection such as articles in news stream. However, Kleinberg's modeling of bursts is usually applied only to bursts of keywords but not to those of topics. Considering this fact, we propose how to apply Kleinberg's modeling of bursts to topics estimated by a topic model such as LDA and DTM (dynamic topic model). © 2012 Springer-Verlag Berlin Heidelberg.
  • Thaer M. Dieb, Masaharu Yoshioka, Shinjiro Hara
    Proceedings of the 2012 IIAI International Conference on Advanced Applied Informatics, IIAIAAI 2012 42 - 47 2012 [Refereed][Not invited]
     
    To support development processes for nanodevices, we want to utilize the information related to experiments from nanodevice development research papers. In this paper, we propose a new guideline for constructing a tagged corpus of nanodevice development papers to achieve this goal. We also propose the use of a corpus construction support tool. To speed up the construction process, we propose an approach for automatic extraction of experiment-related information from the papers. © 2012 IEEE.
  • Masaharu Yoshioka, Noriko Kando
    International Conference on Information and Knowledge Management, Proceedings 19 - 20 2012 [Refereed][Not invited]
     
    We propose a novel framework that enables multifaceted analysis of news articles. This system uses semantic annotated information (e.g., person, place) as facets and can be used to construct structured queries for comparing the differences between sets of articles.
  • Masaharu Yoshioka, Noriko Kando
    PROCEEDINGS OF THE FIFTH WORKSHOP ON EXPLOITING SEMANTIC ANNOTATIONS IN INFORMATION RETRIEVAL 19 - + 2012 [Refereed][Not invited]
     
    We propose a novel framework that enables multifaceted analysis of news articles. This system uses semantic annotated information (e. g., person, place) as facets and can be used to construct structured queries for comparing the differences between sets of articles.
  • 吉岡 真治
    情報処理学会論文誌 情報処理学会 52 (12) 3423 - 3434 1882-7764 2011/12/15 [Not refereed][Not invited]
     
    In order to make a good question answering (QA) system based on text database, it is better to have a good information retrieval (IR) system. To make a good IR system for QA about particular named entities, it is better to use Boolean IR model by using appropriate Boolean query. In this paper, we propose to use ABRIR (Appropriate Boolean query Reformulation for Information Retrieval) that constructs appropriate Boolean query by using list of synonyms and variation of Japanese katakana description. Evaluation results shows that ABRIR works effectively for the task of IR for QA (NTCIR-8 Geotime).
  • Masaharu Yoshioka
    IPSJ Journal 一般社団法人情報処理学会 52 (12) 3423 - 3434 0387-5806 2011/12 [Refereed][Not invited]
     
    In order to make a good question answering (QA) system based on text database, it is better to have a good information retrieval (IR) system. To make a good IR system for QA about particular named entities, it is better to use Boolean IR model by using appropriate Boolean query. In this paper, we propose to use ABRIR (Appropriate Boolean query Reformulation for Information Retrieval) that constructs appropriate Boolean query by using list of synonyms and variation of Japanese katakana description. Evaluation results shows that ABRIR works effectively for the task of IR for QA (NTCIR-8 Geoti...
  • Masaharu Yoshioka, Noriko Kando, Yohei Seki
    Proceedings of the 4th International Workshop on Evaluating Information Access, EVIA 2011, National Center of Sciences, Tokyo, Japan, December 6, 2011 National Institute of Informatics (NII) 20 - 23 2011 [Refereed][Not invited]
  • Fredric C. Gey, Ray R. Larson, Jorge Machado, Masaharu Yoshioka
    Proceedings of the 4th International Workshop on Evaluating Information Access, EVIA 2011, National Center of Sciences, Tokyo, Japan, December 6, 2011 National Institute of Informatics (NII) 9 - 13 2011 [Refereed][Not invited]
  • Thaer M. Dieb, Masaharu Yoshioka, Shinjiro Hara
    Proceedings - 2011 IEEE International Conference on Granular Computing, GrC 2011 167 - 170 2011 [Refereed][Not invited]
     
    Nanodevices development process is not well systematized. In order to support this development process, we have been working on an experimental record management system that aims at analyzing previous manufacturing experiments to provide an insight on future manufacturing. However, we found that experimental records do not have enough information for detailed analysis. In this paper, we propose an approach to extract extra information (metadata) from Nanodevices development papers that can enhance the analysis of experimental results. © 2011 IEEE.
  • Masaharu Yoshioka
    Proceedings of the 9th NTCIR Workshop Meeting on Evaluation of Information Access Technologies: Information Retrieval, Question Answering and Cross-Lingual Information Access, NTCIR-9, National Center of Sciences, Tokyo, Japan, December 6-9, 2011 National Institute of Informatics (NII) 2011 [Refereed][Not invited]
  • Fredric C. Gey, Ray R. Larson, Jorge Machado, Masaharu Yoshioka
    Proceedings of the 9th NTCIR Workshop Meeting on Evaluation of Information Access Technologies: Information Retrieval, Question Answering and Cross-Lingual Information Access, NTCIR-9, National Center of Sciences, Tokyo, Japan, December 6-9, 2011 National Institute of Informatics (NII) 2011 [Refereed][Not invited]
  • Masaharu Yoshioka, Katsuhiro Tomioka, Shinjiroh Hara, Takashi Fukui
    iiWAS2010 - 12th International Conference on Information Integration and Web-Based Applications and Services 871 - 874 2010 [Refereed][Not invited]
     
    We are developing a framework for knowledge creation in nanodevice development, based on collaboration between nanodevice engineers and computer science researchers. Development of nanodevices requires a variety of knowledge some of this knowledge is tacit, based on the user's experience. Therefore, it is difficult to become a good engineer in this development process. We propose the concept of "Evidence-based experiment planning" and develop a process for supporting experiment planning in nanodevice development. This system applies knowledge discovery techniques to records of previous experiments to extract experienced engineers' tacit knowledge. Copyright 2010 ACM.
  • Yusuke Abe, Takehito Utsuro, Yasuhide Kawada, Tomohiro Fukuhara, Noriko Kando, Masaharu Yoshioka, Hiroshi Nakagawa, Yoji Kiyota, Masatoshi Tsuchiya
    ACTIVE MEDIA TECHNOLOGY 6335 498 - + 0302-9743 2010 [Refereed][Not invited]
     
    Among other domains and topics on which some issues are frequently argued in the blogosphere, the domain of crime is one of the most seriously discussed by various kinds of bloggers. Such information on crimes in blogs is especially valuable for those who are not familiar with tips for preventing being victimized. This paper proposes a framework of extracting people's concerns and reports on crimes in their own blogs. In this framework, we solve two tasks. In the first task, we focus on experts in crime domain and address the issue of extracting concerns on crimes such as tips for preventing being victimized. In the evaluation of this first task, we show that we successfully rank expert bloggers high in the results of blog feed retrieval. In the second task, on the other hand, we focus on victims of criminal acts, and address the issue of extracting reports on being victimized In the evaluation of this second task, we show that we outperform blog post ranking based on the search engine API by incorporating dependency relations for identifying victims' blog posts.
  • Masaharu Yoshioka
    INFORMATION RETRIEVAL TECHNOLOGY 6458 588 - 598 0302-9743 2010 [Refereed][Not invited]
     
    To make a good question answering (QA) system based on a text database, it is preferable to have a good information retrieval (IR) system that can find appropriately relevant document sets for a given query. To make a good IR system for QA about particular named entities (NEs), it is preferable to use a Boolean IR model that uses appropriate Boolean queries with the NE information. In this paper, we propose to use appropriate Boolean query reformulation for information retrieval (ABRIR) for this problem. In this system, an appropriate list of synonyms and variations of Japanese katakana descriptions of a given query are used to construct the Boolean query. Evaluation results show that ABRIR works effectively for the IR. task in QA.
  • Hiroyuki Nakasaki, Yusuke Abe, Takehito Utsuro, Yasuhide Kawada, Tomohiro Fukuhara, Noriko Kando, Masaharu Yoshioka, Hiroshi Nakagawa, Yoji Kiyota
    Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering 40 315 - 320 1867-8211 2010 [Refereed][Not invited]
     
    Among other domains and topics on which some issues are frequently argued in the blogosphere, the domain of crime is one of the most seriously discussed by various kinds of bloggers. Such information on crimes in blogs is especially valuable for outsiders from abroad who are not familiar with cultures and crimes in foreign countries. This paper proposes a framework of cross-lingually analyzing people's concerns, reports, and experiences on crimes in their own blogs. In the retrieval of blog feeds/posts, we take two approaches, focusing on various types of bloggers such as experts in the crime domain and victims of criminal acts. © Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering 2010.
  • Masaharu Yoshioka
    Proceedings of the 8th NTCIR Workshop Meeting on Evaluation of Information Access Technologies: Information Retrieval, Question Answering and Cross-Lingual Information Access, NTCIR-8, National Center of Sciences, Tokyo, Japan, June 15-18, 2010 National Institute of Informatics (NII) 154 - 158 2010 [Refereed][Not invited]
  • NSContrast: An Exploratory News Article Analysis System that Characterizes the Differences between News Sites
    Masaharu Yoshioka
    SIGIR2009 Workshop on Information Access in a Multilingual World 25 - 29 2009/07 [Refereed][Not invited]
  • NSContrast: Characterizing the Difference of Interest among Multiple News Sites
    Masaharu Yoshioka
    INTERACT 2009 Workshop on Culture and Technologies for Social Interaction http://sites.google.com/site/crosscultureworkshopinteract09/position-papers-1/10-Yoshioka.pdf-29  2009 [Refereed][Not invited]
  • TORISAWA Kentaro, NAKAGAWA Hiroshi, KUROHASHI Sadao, INUI Kentaro, YOSHIOKA Masaharu, FUJII Atsushi, KITSUREGAWA Masaru
    IPSJ Magazine 一般社団法人情報処理学会 49 (8) 890 - 896 0447-8053 2008/08/15 [Not refereed][Not invited]
  • Kiyotaka Takahashi, Aki Sugiyama, Yoshiki Shimomura, Takeshi Tateyama, Ryosuke Chiba, Masaharu Yoshioka, Hideaki Takeda
    Proceedings of the 10th International Conference on Information Integration and Web-based Applications and Services, iiWAS 2008 18th 575 - 578 2424-3078 2008 [Refereed][Not invited]
     
    Recently, designers have been required solving comprehensive problem becoming greater and more complicated. In relation to this background, we have proposed Universal Abduction Studio (UAS) a computer environment that synthetically supports creative design. However, it is difficult for a designer to manually acquire multiple domain knowledge required for UAS. Therefore, we propose a Web-based knowledge database construction method for supporting design by UAS in this paper. © 2008 ACM.
  • IR Interface for Contrasting Multiple News Sites
    Masaharu Yoshioka
    Proceedings of the Asia Information Retrieval Symposium 2008 - (-) 516 - 521 2008/01 [Refereed][Not invited]
  • Masaharu Yoshioka
    SITIS 2008: 4TH INTERNATIONAL CONFERENCE ON SIGNAL IMAGE TECHNOLOGY AND INTERNET BASED SYSTEMS, PROCEEDINGS 45 - 51 2008 [Refereed][Not invited]
     
    Today, there is access to large numbers of news sites in different countries, and there are several experimental systems, such as Newsblaster and NewsExplorer that integrate news articles about a particular event from multiple news sites. These systems enable a good understanding of particular events by using multiple news sites, but they ignore the characteristics of the various news sites. To characterize the differences between news sites, the News Site Contrast (NSContrast) system has been proposed. This system analyzes multiple news sites using the concept of contrast-set mining. However NSContrast has only limited analysis functions and is not mature enough for evaluation via user experiments. Therefore, in this paper, we analyze contrast set mining results for NSContrast, aiming to understand the requirements for extracting useful information that also reflects the interests of different countries. Based on this analysis, a new NSContrast is introduced and applied to a news article database constructed from multiple news sites in different countries for user experimentation.
  • Kiyotaka Takahashi, Aki Sugiyama, Yoshiki Shimomura, Takeshi Tateyama, Ryosuke Chiba, Masaharu Yoshioka, Hideaki Takeda
    PRACTICAL ASPECTS OF KNOWLEDGE MANAGEMENT, PROCEEDINGS 5345 173 - + 0302-9743 2008 [Refereed][Not invited]
     
    In recent years, comprehensive problem solving by artefact designers has been required as demands related to artefacts become greater and more complicated. In relation to this background, we have proposed Universal Abduction Studio (UAS), a computer environment that synthetically supports a creative design. However, in general, it is difficult for a designer to acquire multiple domain knowledge because this knowledge is accumulated manually. This paper proposes a Web-based knowledge database construction method for supporting design using UAS.
  • 吉岡真治
    ヒューマンインタフェース学会誌 9 (4) 269-274  1344-7254 2007/11 [Not refereed][Invited]
  • TAKEDA Hideaki, SHIMOMURA Yoshiki, YOSHIOKA Masaharu
    Journal of Japan Society for Design Engineering 日本設計工学会 42 (10) 558 - 564 0919-2948 2007/10/05 [Not refereed][Not invited]
  • Hideaki Takeda, Yutaka Fujimoto, Masaharu Yoshioka, Yoshiki Shimomuraz, Kengo Morimoto, Wataru Oniki
    NEW FRONTIERS IN ARTIFICIAL INTELLIGENCE 3609 (-) 216 - + 0302-9743 2007 [Refereed][Not invited]
     
    This paper describes how to add tags to design documents in order to extract knowledge from information for intelligent design support. Our project called Universal Abduction Studio (UAS) aims to build a new design support system that supports conceptual design by dynamically integrating knowledge in different design domains. This paper focuses on knowledge description form which can be used to capture knowledge from text-based information and then be used for inference for creative design. We propose so-called design knowledge document containing both human-readable texts and machine-readable knowledge such as propositions and rules.
  • Masaharu Yoshioka
    INFORMATION RETRIEVAL TECHNOLOGY, PROCEEDINGS 4182 390 - 403 0302-9743 2006 [Refereed][Not invited]
     
    Query term expansion is an important technique for achieving higher retrieval performance. However, since many factors affects the quality of this technique, it is difficult to evaluate this technique in isolation. Feature quantities that characterize the quality of the initial query are defined in this study for evaluating topic difficulties from the viewpoint of query term expansion. I also briefly review the result of the NTCIR-5 query term expansion subtask that uses these quantities for evaluating the effectiveness of the query term expansion techniques. I also describe detailed analysis results on the effect of query term expansion based on topic-by-topic analysis.
  • SHIMOMURA Yoshiki, YOSHIOKA Masaharu, TAKEDA Hideaki, TOMIYAMA Tetsuo
    TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C 一般社団法人 日本機械学会 72 (713) 274 - 281 0387-5024 2006 [Not refereed][Not invited]
     
    This paper describes a new design support system that supports conceptual or creative design by dynamically integrating knowledge in different design domains. We argue that abduction for integrating theories can be a basic principle to formalize such design processes. Based on this principle, we propose Universal Abduction Studio, a design environment in which designers combine different theories to arrive at better design. In this new approach to computational support of conceptual design, the system should offer various types of abductive reasoning from which designers can select an interesting design method. We also discuss technologies to implement UAS and in this paper we propose to use analogical reasoning as abductive reasoning to discover relationships between knowledge from different sources. We demonstrate that a prototype system can discover a new idea in a design example taken from a real design activity.
  • YOSHIOKA Masaharu, HARAGUCHI Makoto
    Transactions of the Japanese Society for Artificial Intelligence 一般社団法人 人工知能学会 20 (4) 270 - 280 1346-0714 2005/11/01 [Not refereed][Not invited]
     
    It is not easy for a user of an Information Retrieval (IR) system to select an appropriate keyword set to represent his or her specific information need. Therefore, many IR systems can modify keyword sets by estimating the user's particular requirement. Even though such IR systems have better retrieval performance, the complicated estimation process entailed by a large number of keywords makes it difficult for a user to understand how the system behaves. Therefore, we used a thesaurus for query expansion. To select an appropriate keyword set, we proposed two concepts: ``adaptive generalization'', which estimates an appropriate generalization level of the given keywords by using relevant document information, and ``purpose-oriented concept structure modification'', which selects relevant keywords from a predefined synonym set in a thesaurus. Because query expansion based on a thesaurus aims to find new keywords that are complementary to the initial keywords, we proposed to use this method to construct a Boolean query formula to represent the user's information need. We proposed a new IR system called ``appropriate Boolean query reformulation for IR with adaptive generalization'' (ABRIR-AG) to support Boolean query formation. In ABRIR-AG, we reformulated a user-given Boolean query by using a small number of relevant documents. Finally, to evaluate its effectiveness, we evaluated ABRIR-AG by using a large-scale test collection containing WWW documents.
  • An Architecture for Designers' Support Systems with Knowledge-embedded Documents
    Hideaki Takeda, Masaharu Yoshioka, Yoshiki Shimomura, Yutaka Fujimoto, Kengo Morimoto, Wataru Oniki
    Proceedings of the 15th International Conference on Engineering Design -ICED05- (CD-ROM) 2005 [Refereed][Not invited]
  • M Yoshioka, M Haraguchi
    International Workshop on Challenges in Web Information Retrieval and Integration, Proceedings 145 - 150 2005 [Refereed][Not invited]
     
    When implementing an IR system that can support comprehensive searches from a wide variety of documents, it is crucial to have a mechanism for selecting appropriate query formulations. For this purpose, many IR systems can modify query terms by estimating the user's information need. However because modified query terms are usually represented in a complicated form, it is difficult to judge how appropriate they are. In this research, we assume all relevant documents should contain words that correspond to concepts that the user would like to retrieve. Therefore, a Boolean query formula is one good way to represent a user's information need with high readability. To find the required concepts in a relevant document set, we have previously proposed a new IR system that calculates appropriate generalized concepts to represent the user's need. In this study, we implement a user interface that supports reformulation of IR queries by using abstract concepts.
  • Masaharu Yoshioka, Makoto Haraguchi
    ACM Transactions on Asian Language Information Processing 4 (3) 340 - 356 1530-0226 2005 [Refereed][Not invited]
     
    Even though a Boolean query can express the information need precisely enough to select relevant documents, it is not easy to construct an appropriate Boolean query that covers all relevant documents. To utilize a Boolean query effectively, a mechanism to retrieve as many as possible relevant documents is therefore required. In accordance with this requirement, we propose a method for modifying a given Boolean query by using information from a relevant document set. The retrieval results, however, may deteriorate if some important query terms are removed by this reformulation. A further mechanism is thus required in order to use other query terms that are useful for finding more relevant documents, but are not strictly required in relevant documents. To meet this requirement, we propose a new method that combines the probabilistic IR and the Boolean IR models. We also introduce a new IR system - called appropriate Boolean query reformulation for information retrieval (ABRIR) - based on these two methods and the Okapi system. ABRIR uses both a word index and a phrase index formed from combinations of two adjacent noun words. The effectiveness of these two methods was confirmed according to the NTCIR-4 Web test collection. © 2005 ACM.
  • M Yoshioka
    INFORMATION RETRIEVAL TECHNOLOGY, PROCEEDINGS 3689 647 - 652 0302-9743 2005 [Refereed][Not invited]
     
    Query expansion is an important technique for achieving higher retrieval performance. However, a good framework for evaluating this technique in isolation does not, exist because the effect of query expansion depends on the quality of the initial query. Feature quantities that characterize the quality of the initial query are defined in this study. Correlations between these quantities and retrieval performance were analyzed using the NTCIR-4 web test collection.
  • M Yoshioka, Y Umeda, H Takeda, Y Shimomura, Y Nomaguchi, T Tomiyama
    ADVANCED ENGINEERING INFORMATICS 18 (2) 95 - 113 1474-0346 2004/04 [Refereed][Not invited]
     
    Knowledge intensive engineering aims at flexible applications of a variety of product life cycle knowledge. much as design, manufacturing. operations, maintenance, and recycling. Many engineering domain theories are organized and embedded within CAD and CAE tools and engineering activities can be formalized as modeling operations to them. Since most of domain theories deal with the physical world and can be associated with physical concepts. a physical concept ontology can form a common ontology to integrate engineering models that are formed based on domain theories. This paper reports a physical ontology-based support system for knowledge intensive engineering called Knowledge intensive Engineering Framework (KIEF) to integrate multiple engineering models and to allow more flexible use of them. First the paper describes the physical ontology as the core of KIEF and an ontology-based reasoning system called a pluggable metamodel mechanism, to integrate and maintain relationships among these models. The pluggable metamodel mechanism uses a metamodel that represents the designer's mental model about a design object as a concept network model. The designer builds and decomposes a functional hierarchy from functional specifications with an FBS (Function-Behavior-State) modeler. He/She then maps the functional hierarchy into a metamodel using physical features that are building blocks for conceptual design. Then, the pluggable metamodel mechanism enriches the information contained in the metamodel by using causal dependency knowledge about the physical world and by budding and analzing various engineering models. We demonstrate the power of KIEF by illustrating a design case performed on KIEF. (C) 2004 Published by Elsevier Ltd.
  • M Yoshioka, M Haraguchi
    8TH WORLD MULTI-CONFERENCE ON SYSTEMICS, CYBERNETICS AND INFORMATICS, VOL I, PROCEEDINGS I 226 - 230 2004 [Not refereed][Not invited]
     
    We can use various types of information retrieval (IR) system to access large quantities of text data. These systems usually request users to input keywords to represent their retrieval intention. Since, it is difficult for a user to select appropriate keywords, many IR systems have the capability of modifying keywords by estimating the user's intention. However because modified keywords are usually represented in a complicated form, it is difficult to judge how appropriate they are. In this research, we assume all relevant documents should have words that correspond to concepts that the user would like to retrieve. Therefore, a boolean query formula is one good way to represent user's intent with higher readability. In order to find out these concepts in a relevant document set, we have already proposed a new IR system that calculates appropriate generalized concept to represent user's intention. In this study, we implement a user interface that supports to clarify IR query by using abstract concepts.
  • M Yoshioka, M Haraguchi, A Mizoe
    INTUITIVE HUMAN INTERFACES FOR ORGANIZING AND ACCESSING INTELLECTUAL ASSETS 3359 243 - 255 0302-9743 2004 [Refereed][Not invited]
     
    In order to construct story databases, it is crucial to have an effective index that represents the plot and event sequences in a document. For this purpose, we have already proposed a method using the concept of maximal analogy to represent a generalized event sequence of documents with a maximal set of events. However, it is expensive to calculate a maximal analogy from documents with a large number of sentences. Therefore, in this paper, we propose an efficient algorithm to generate a maximal analogy, based on graph theory, and we confirm its effectiveness experimentally. We also discuss how to use a maximal analogy as an index for a story database, and outline our future plans.
  • Universal abduction studio -- proposal of a design support environment for creative thinking in design --
    Hideaki Takeda, Hiromitsu Sakai, Yutaka Nomaguchi, Masaharu Yoshioka, Yoshiki Shimomura, Tetsuo Tomiyama
    A. Folkman, K. Gralen, M. Norell, and U. Sellgren, editors, The Fourteenth International Conference on Engineering Design (ICED 03), Stockholm, August 2003 2003/08 [Refereed][Not invited]
  • Masaharu Yoshioka, Makoto Haraguchi
    Proc. of Asian Association for Lexicography' 03 (ASIALEX) 461 - 466 2003 [Not refereed][Not invited]
  • M Haraguchi, M Yotsutani, M Yoshioka
    7TH WORLD MULTICONFERENCE ON SYSTEMICS, CYBERNETICS AND INFORMATICS, VOL V, PROCEEDINGS V 213 - 216 2003 [Not refereed][Not invited]
     
    To classify a document set from various viewpoints, we extract a possible abstract document from given similar documents. The abstract one is a common generalization of the original documents. Although we can present an efficient algorithm to find it under some pruning rule, it is generally very hard to perform such a task of extracting abstract documents. It is therefore necessary to reduce the numbers of events appearing in documents. For this purpose, we regard one document as a network of events, present a pseudo-PageRank model for calculating the importance of events, analyze and compare it with a standard text summarization method, and finally reports the present approach to improve the pseudo-PageRank.
  • Masaharu Yoshioka, Yasuhiro Shamoto
    Volume 1: 23rd Computers and Information in Engineering Conference, Parts A and B 2003/01/01 [Refereed]
     
    In engineering activities various types of problems have to be solved. Most of these problems are routine problems and some are complicated problems for which it is difficult to formalize all the problem-solving processes. We propose a knowledge management system for supporting these problem solving processes. One example is using information applicable to new and complicated problems that has resulted from solving similar problems in the past. In addition, in such complicated problem solving processes, most of the decomposed sub-problems can be handled as routine problems. In this paper, we propose use of formalized knowledge and document information that represents the history of problem-solving processes to support the solution of both types of problem. For that purpose, we propose the Documentation Management system for Problem Solving (DMaPS), which supports a user solving a problem by document processing. In this support process, the system retrieves related documents for the user’s problems using formalized knowledge and information retrieval techniques. After retrieving related documents, the user modifies them to fit their problem. The modified information is associated with the formalized knowledge and is stored as documents for future reuse.
  • Abduction for Creative Design
    Tetsuo Tomiyama, Hideaki Takeda, Masaharu Yoshioka, Yoshiki Shimomura
    Computational Synthesis: From Basic Building Blocks to High Level Functionality Papers from the 2003 AAAI Symposium 239 - 246 2003 [Not refereed][Not invited]
  • Tetsuo Tomiyama, Hideaki Takeda, Masaharu Yoshioka, Yoshiki Shimomura
    Proceedings of the ASME Design Engineering Technical Conference 3 543 - 552 2003 [Refereed][Not invited]
     
    While abduction is considered crucial for design in general, this paper focuses on the role of abduction to integrate knowledge assuming that creative design can come from innovative combination of existing knowledge. Based on Schurz's classification of abductive reasoning, the paper identifies that abduction for integrating theories can be performed by a special type of abduction called second order existential abduction. The paper then analyzes refrigerator design cases to understand how knowledge is used and shows that abduction is indeed central to design. It also discusses that knowledge structure is a key concept in abduction for integration.
  • Oyama, K., Kageura, K., K, o, N., Kimura, M., Maruyama, K., Yoshioka, M., Takahashi, K.
    Systems and Computers in Japan 34 (6) 44 - 56 2003/01 [Refereed][Not invited]
  • TAKEUCHI Koichi, UCHIYAMA Kiyoko, YOSHIOKA Masaharu, KAGEURA Kyo, KOYAMA Teruo
    Transactions of Information Processing Society of Japan 一般社団法人情報処理学会 43 (5) 1446 - 1456 1882-7764 2002/05/15 [Not refereed][Not invited]
     
    In this paper, we describe a principled approach for analyzing relations between constituent words of compound nouns, specifically those whose heads are deverbal nouns, based on the classification of deverbal nouns by their lexical conceptual structure (LCS) and the classification of nouns in modifier position vis-a-vis LGS of head deverbal nouns. There are two kinds of relations of compounds with head deverbal nouns. The one is that a modifier noun becomes an argument of deverbal head and the other is that a modifier becomes an adjunct. It is an important starting point for analyzing relations to disambiguate the two kinds of relations. Through the qualitative analysis of the data and the experimental evaluation of 1231 compound nouns, we show that the use of LCS as the theoretical basis is very promising for constructing compound analyzer.
  • OGAWA Yasushi, SASAKI Yutaka, MASUYAMA Shigeru, MURATA Masaki, YOSHIOKA Masaharu, Yasushi Ogawa, Yutaka Sasaki, Shigeru Masuyama, Masaki Murata, Masaharu Yoshioka
    Journal of Japanese Society for Artificial Intelligence 人工知能学会 17 (3) 306 - 311 0912-8085 2002/05/01 [Not refereed][Not invited]
  • 人工物工学研究の展開 第18報 -Universal Abduction Studioの提案-
    坂井宏充, 野間口大, 吉岡真治, 武田英明, 下村芳樹, 冨山哲男
    日本機械学会第12回設計工学・システム部門講演会講演論文集 2002/04 [Not refereed][Not invited]
  • Document-based Design Process Knowledge Management for Knowledge Intensive Engineering
    Y. Nomaguchi, M. Yoshioka, T. Tomiyama
    From Knowledge Intensive CAD to Knowledge Intensive Engineering Kluwer Academic Publishers 2002/04 [Refereed][Not invited]
  • KURIYAMA KAZUKO, YOSHIOKA MASAHARU, KANDO NORIKO
    情報処理学会論文誌データベース（TOD） 一般社団法人情報処理学会 43 (2) 48 - 59 1882-7799 2002/03/15 [Not refereed][Not invited]
     
    The purposes of this study are to examine whether there is an effect on the relative evaluation of the IR systems using the relevance judgments of the test collection NTCIR-2 made by the pooling method and additional interactive searches, and to investigate whether the NTCIR-2 is effective for evaluating the IR systems, the search results of which were not used for the pooling. We carried out experiments using different lists of relevance judgments and search results submitted for the test of the 2nd NTCIR Workshop. First, we evaluated the search results using the list of the final relevance judgments F of NTCIR-2 and F - I, that is, the F without the unique relevant documents found by the additional interactive searches I. Second, we made pools from the search results for each of the sub-tasks and evaluated the search results using the pools as lists of relevance judgments. Third, we evaluated the search results using F - S, that is, the F without the unique relevant documents found by an IR system S in order to simulate evaluation of the IR system which was not used for the pooling. Almost the same rankings of the search results were produced by using the pools as lists of relevance judgments for system evaluation. When the search results by an JR system were not used for the pooling, there is a very little effect on evaluation of the system itself and relative evaluation among it and other systems. Therefore our results verified the reliability of test collection as an evaluation tool, which was based on pooling method.
  • K Takeuchi, K Uchiyama, M Yoshioka, K Kageura, T Koyama
    2001 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS, VOLS 1-5 904 - 909 1062-922X 2002 [Refereed][Not invited]
     
    This paper proposes a method to describe conceptual meanings of Japanese deverbal nouns for the purpose of analysing Japanese compound nouns. A lexical meaning of each constituent word is very important since it deeply influences a word formation of compound words. Therefore we intend to use the lexical meanings of constituent words to build a Japanese compound noun analyser. As a first step to describe lexical meanings of words, we have established a simple semantic structure based on lexical conceptual structure (we refer to the structure we establish by TLCS1 in this paper). After the analysing 250 deverbal nouns that appeared in technical terms, we found that the only 14 types of TLCSes cover various deverbal nouns. We will explain in this paper the types of TLCS and also show the procedure of assigning TLCSes to deverbals. In order to evaluate the efficacy of TLCS, we made use of the accuracy of simple compound noun analyser for analysing technical terms. The result was very good, which shows that our proposed method is very promising.
  • Yutaka Nomaguchi, Masaharu Yoshioka, Tetsuo Tomiyama
    Transactions of the Japanese Society for Artificial Intelligence 17 (1) 104 - 113 1346-0714 2002 [Not refereed][Not invited]
     
    In this paper, we propose a fundamental idea of a new CAD mechanism to facilitate design knowledge management. This mechanism encourages a designer to externalise his/her knowledge during a design process and facilitates sharing and reuse of such externalised design knowledge in later stages. We also describe the implementation of this idea called DDMS (Design Documentation Management System). DDMS works as a front end to KIEF (Knowledge Intensive Engineering Framework), which we have been developing. We also illustrate an example of machining tool design to demonstrate the features of DDMS.
  • M Haraguchi, S Nakano, M Yoshioka
    DISCOVERY SCIENCE, PROCEEDINGS 2534 324 - 331 0302-9743 2002 [Not refereed][Not invited]
     
    Given two documents in the form of texts, we present a notion of maximal analogy representing a generalized event sequence of documents with a maximal set of events. They are intended to be used as extended indices of documents to automatically organize a document database from various viewpoints. The maximal analogy is defined so as to satisfy a certain consistency condition and a cost condition. Under the consistency condition, a term in an event sequence is generalized to more abstract term independently of its occurrence positions. The cost condition is introduced so that meaningless similarities between documents are never concluded. As the cost function is monotone, we can present an optimized bottom-up search procedure to discover a maximal analogy under an upper bound of cost. We also show some experimental results based on which we discuss a future plan.
  • A Knowledge Operation model of Synthesis
    TOMIYAMA Tetsuo, YOSHIOKA Masaharu, TSUMAYA Akira
    Engineering Design Synthesis - Understanding, Approaches, and Tools, Springer-Verlag pp.67-90 2002/01 [Not refereed][Not invited]
  • 野間口大, 吉岡真治, 冨山哲男
    人工知能学会論文誌 社団法人 人工知能学会 17 104 - 113 1346-0714 2002 [Refereed][Not invited]
     
    In this paper, we propose a fundamental idea of a new CAD mechanism to facilitate design knowledge management. This mechanism encourages a designer to externalise his/her knowledge during a design process and facilitates sharing and reuse of such externalised design knowledge in later stages. We also describe the implementation of this idea called DDMS (Design Documentation Management System). DDMS works as a front end to KIEF (Knowledge Intensive Engineering Framework), which we have been developing. We also illustrate an example of machining tool design to demonstrate the features of DDMS.
  • Masaharu Yoshioka, Yutaka Nomaguchi, Tetsuo Tomiyama
    Volume 2B: 27th Design Automation Conference 2001/09/09 [Refereed][Not invited]
     
    Abstract We propose a new integrated computational environment, called a Knowledge Intensive Engineering Framework (KIEF), to support design object modeling, during engineering design process that requires a variety of design object models, such as a geometric model, a control model, and a finite element model. KIEF is capable of guiding designers with design process knowledge based on a model of synthesis, of integrating multiple design object models, and of maintaining consistency among these models. The design process knowledge is a result of a research project we conduct to establish a model of synthesis. This model formalizes design process knowledge in a design process as operations to a multiple model-based reasoning system, such as KIEF, and from a language of synthesis. We describe a prototype system that incorporates design process knowledge to KIEF and illustrate an example design on it.
  • YOSHIOKA Masaharu, TOMIYAMA Tetsuo
    Journal of the Japan Society for Precision Engineering 公益社団法人精密工学会 67 (9) 1490 - 1496 0912-0289 2001/09/05 [Not refereed][Not invited]
     
    In a design process, a designer uses various kinds of design object knowledge, such as physical laws, mechanism library, and manufacturing knowledge. Therefore, it is necessary for a CAD system to handle large scale knowledge bases to support this process. However, it is difficult to use a large scale knowledge base without knowledge about operations of the knowledge itself. This paper proposes a theory of design knowledge operations. First, we analyze the results of design experiments that are records of real design processes. Second, we formalize design knowledge operations based on General Design Theory. Finally we discuss the generality of the proposed formalism of design knowledge operations.
  • 内山 清子, 竹内 孔一, 吉岡 真治
    計量国語学 計量国語学会 23 (1) 1 - 24 0453-4611 2001/06 [Not refereed][Not invited]
  • A Model of Synthesis: A System of Multiple Theories
    TOMIYAMA Tetsuo, HEW Ken P, MURAKAMI Tamotsu, YOSHIOKA Masaharu, TAKEDA Hideaki, WASHIO Takeshi, TSUMAYA Akira, UMEDA Yasushi
    Proceedings of the International CIRP Design Seminar 2001 pp.1-6 2001/06 [Refereed][Not invited]
  • OYAMA Keizo, KAGEURA Kyo, KANDO Noriko, KIMURA Masaru, MARUYAMA Katsumi, YOSHIOKA Masaharu, TAKAHASHI Kazumichi
    The Transactions of the Institute of Electronics,Information and Communication Engineers. 一般社団法人電子情報通信学会 84 (6) 658 - 670 0915-1915 2001/06/01 [Not refereed][Not invited]
     
    インターネットの発展に伴い, WWWを介した本格的な情報検索サービスの堤供が求められているが, 既存のWWW情報検索システムでは, 検索機能やデータベースの構造・規模などに制約が多い.本論文の情報検索システムは実際の情報検索サービスで運用されているものであり, 大規模で複雑な構造をもつテキスト型の学術情報データベースに対し, 高度な検索機能を, ラインモードとWWWインタフェースにより提供している.本論文では, まず情報検索機能の概要と設計について, 特に集合演算, シソーラス検索, 階層構造検索, 複数データベースの統合検索などの特徴的な機能について述べている.次に, その実現のための実装技術として, 検索エンジンの特長, セッション管理方式, 大規模データベースの扱い, サーバ及びプロセスの構成などを示し, 現状のシステムについての運用上の評価を行い, 最後に今後の課題と関連研究について議論している.
  • 設計シミュレータを用いた設計文書作成支援
    野間口大, 冨山哲男, 吉岡真治
    2001年度精密工学会春季大会学術講演会講演論文集 2001/04 [Not refereed][Not invited]
  • H Takeda, M Yoshioka, T Tomiyama
    DESIGN RESEARCH - THEORIES, METHODOLOGIES, AND PRODUCT MODELLING 307 - 314 2001 [Not refereed][Not invited]
  • Y Umemori, S Kondoh, Y Umeda, Y Shimomura, M Yoshioka
    SECOND INTERNATIONAL SYMPOSIUM ON ENVIRONMENTALLY CONSCIOUS DESIGN AND INVERSE MANUFACTURING, PROCEEDINGS 87 - 92 2001 [Not refereed][Not invited]
     
    Conventional products tend to be thrown away because of functional obsoleteness before reaching the end of physical life of the products, This is one of the main causes of the mass disposal problem. In order to solve this problem, this paper proposes a design methodology for upgradable products, Especially, this paper discusses a method dealing with uncertainty caused by long-term planning in the design methodology for upgradability Namely, the proposed method deals with uncertainty as ranges of parameter values, and derives a design solution that realizes required upgrades, while adapting to estimated uncertainty. This method makes the products robust and tolerant against the uncertainty and, therefore, makes design for upgradability more feasible.
  • A Tsumaya, Y Nomaguchi, M Yoshioka, H Takeda, T Murakami, T Tomiyama
    DESIGN RESEARCH - THEORIES, METHODOLOGIES, AND PRODUCT MODELLING 229 - 236 2001 [Not refereed][Not invited]
  • A Model of Synthesis from the Viewpoint of Design Knowledge Operations
    TOMIYAMA Tetsuo, YOSHIOKA Masaharu, TSUMAYA Akira
    Proceedings of the 1st International Conference on Mechanocal Engineering (ICME2000) CD-ROM (6 pages) 2000/11 [Refereed][Not invited]
  • IR/IE/Summarisation Evaluation Projects in Japan
    Kageura,K, Fukushima, T, Kando, N, Okumura, M, Sekine, S, Kuriyama, K, Takeuchi, K, Yoshioka, M, Koyama, T, Isahara, H
    Proceedings of LREC 2000 Workshop on Using Evaluation within HLT Programs, Athens, Greece, May, 2000 19 - 22 2000/05 [Refereed][Not invited]
  • Kyo Kageura, Masaharu Yoshioka, Koichi Takeuchi, Teruo Koyama, Keita Tsuji, Fuyuki Yoshikane
    Terminology 6 (2) 151 - 173 1569-9994 2000 [Refereed][Not invited]
     
    This article provides basic background information on the articles included in this special issue on Japanese term extraction, by (i) clarifying the basic background of research into automatic term recognition, (ii) explaining briefl y the ‘contest’-style workshop we organised in 1999, and (iii) briefl y summarising the ATR methodologies proposed in the articles, and positioning their ideas, philosophies and methodologies within ATR from a unifi ed perspective. Through this information, we intend to consolidate the contributions of the NTCIR TMREC workshop, and, hopefully, clarify a basic framework for discussion which different researchers can use to constructively communicate with each other about automatic term extraction and beyond. © 2000 John Benjamins Publishing Co.
  • KIMURA Masaru, YOSHIOKA Masaharu, KANDO Noriko, KAGEURA Kyo, OYAMA Keizo
    The Journal of Information Science and Technology Association 一般社団法人 情報科学技術協会 50 (1) 30 - 36 0913-3801 2000 [Not refereed][Not invited]
     
    This paper describes the transformation of NACSIS information retrieval service (NACSIS-IR). NACSIS has provided database search service called NACSIS-IR since 1987. NACSIS-IR plans to start new information retrieval service based on fulltext search engine in January 2000. The authors introduce structure of databases, search functions, WWW-based user interfaces, and other features of new NACSIS-IR.
  • SEKIYA Takayuki, YOSHIOKA Masaharu, TOMIYAMA Tetsuo, Takayuki Sekiya, Masaharu Yoshioka, Tetsuo Tomiyama
    Journal of Japanese Society for Artificial Intelligence 人工知能学会 14 (6) 1051 - 1060 0912-8085 1999/11/01 [Not refereed][Not invited]
     
    During design, a designer uses various computational tools, such as a geometric modeling system, analysis tools, and databases. To support these design processes, a system that can integrate such computational tools is required. For this purpose, we have been developing the Knowledge Intensive Engineering Framework (KIEF) system. KIEF enables the designer to build, evaluate, and modify design models on the system by providing him/her with a wide variety of engineering knowledge. A common ontology plays an important role in integrating tools and representing engineering knowledge. This paper describes the structure of the ontology in KIEF, and how it is used in a modeling processes on KIEF.
  • NOMAGUCHI Y.
    Proceedings of the 1999 Lancaster Internationla Workshop on Engineering Design, CACD'99 135 - 163 1999/05 [Refereed][Not invited]
  • 設計文書からの知識の構造化の手法の提案
    野間口大, 吉岡真治, 冨山哲男
    日本機械学会第9回設計工学・システム部門講演会論文集 1999/04 [Not refereed][Not invited]
  • Kyo Kageura, Masaharu Yoshioka, Koichi Takeuchi, Teruo Koyama, Keita Tsuji, Fuyuki Yoshikane, Maho Okada
    Proceedings of the First NTCIR Workshop on Research in Japanese Text Retrieval and Term Recognition, NTCIR-1, Tokyo, Japan, August 30 - September 1, 1999 National Center for Science Information Systems (NACSIS) 1999 [Refereed][Not invited]
  • Kyo Kageura, Masaharu Yoshioka, Keita Tsuji, Fuyuki Yoshikane, Koichi Takeuchi, Teruo Koyama
    Proceedings of the First NTCIR Workshop on Research in Japanese Text Retrieval and Term Recognition, NTCIR-1, Tokyo, Japan, August 30 - September 1, 1999 National Center for Science Information Systems (NACSIS) 417 - 434 1999 [Refereed][Not invited]
  • Teruo Koyama, Masaharu Yoshioka, Koichi Takeuchi, Kyo Kageura
    Proceedings of the First NTCIR Workshop on Research in Japanese Text Retrieval and Term Recognition, NTCIR-1, Tokyo, Japan, August 30 - September 1, 1999 National Center for Science Information Systems (NACSIS) 1999 [Refereed][Not invited]
  • Koichi Takeuchi, Masaharu Yoshioka, Teruo Koyama, Kyo Kageura
    Proceedings of the First NTCIR Workshop on Research in Japanese Text Retrieval and Term Recognition, NTCIR-1, Tokyo, Japan, August 30 - September 1, 1999 National Center for Science Information Systems (NACSIS) 1999 [Refereed][Not invited]
  • 設計知識に注目した設計の推論モデルの提案(第2報) --設計行為に関する知識の提案--
    武田英明, 吉岡真治, 冨山哲男
    日本機械学会 第9回 設計工学・システム部門講演会 82 - 85 1999/01 [Not refereed][Not invited]
  • Phrase processing methods for Japanese text retrieval
    Kando, N, Kageura, K, Yoshioka, M, Oyama, K
    Proceedings of ACM-SIGIR Workshop on Information Retrieval: Theory into Practice. August 28, 1998, Melbourne, Australia 13 - 19 1998/08/28 [Refereed][Not invited]
  • YOSHIOKA Masaharu, TOMIYAMA Tetsuo, Masaharu Yoshioka, Tetsuo Tomiyama
    Journal of Japanese Society for Artificial Intelligence 人工知能学会 13 (2) 312 - 319 0912-8085 1998/03/01 [Not refereed][Not invited]
     
    In this study, we propose a framework of an integrated modeling environment for design. Since a mechanical engineering design process requires various kinds of design object models, such as a geometric model, a kinematic model, and a finite element model, this environment should maintain consistency among these models. Furthermore, commercial design tools are practically used in design processes, and these must be integrated as well. For dealing with multiple design object models, we have formalized a concept called a metamodel. A metamodel is a model which represents relationships among concepts used in various design object models. This paper proposes a framework of the pluggable metamodel mechanism that allows to plug in existing design tools, to support a modeling on it, and to maintain consistency among them. We describe a prototype system and illustrate an example of ball screw design.
  • Noriko Kando, Kyo Kageura, Masaharu Yoshioka, Keizo Oyama
    SIGIR Forum (ACM Special Interest Group on Information Retrieval) 32 (2) 23 - 28 0163-5840 1998 [Refereed][Not invited]
     
    This paper examines the effectiveness of different phrase identification and weighting methods for Japanese text retrieval in an operational information retrieval (IR) system, called NACSIS-IR. Based on our previous experiments, we used character-based indexing with positional information and word-or phrase-based query processing, which allowed us to implement sophisticated linguistic analysis on large-scale databases while maintaining adequate efficiency. The results of retrieval experiments on a large-scale Japanese test collection showed that the combination of enhanced phrase identification using patterns defined over part-of-speech tags and our algorithms Phrase2 and Phrase5 made a significant positive contribution to retrieval effectiveness. The paper also discusses indexing and phrase processing of Japanese or East Asian languages.
  • Representation of design object based on functional evolution process model
    Yoshiki Shimomura, Masaharu Yoshioka, Hideaki Takeda, Yasushi Umeda, Tetsuo Tomiyama
    Journal of Mechanical Design 120 (2) 221 - 229 1998/01 [Refereed][Not invited]
  • NTCIR : NACSIS Test Collection Project
    Noriko Kando, Teruo Koyama, Keizo Oyama, Kyo Kageura, Masaharu Yoshioka
    Proceedings of the 20th Annual Colloquium of BCS-IRSG 1997/01 [Refereed][Not invited]
  • The modeling of synthesis -from the viewpoint of design knowledge
    T. Tomiyama, T. Murakami, T. Washio, A. Kubota, H. Takeda, T. Kiriyama, Y. Umeda, M. Yoshioka
    WDK 25, Proceedings of the 11th International Conference on Engineering Design in Tampere 199, Tampere, Finland, 1997. 3 97 - 100 1997/01 [Refereed][Not invited]
  • 吉岡真治, 中村雅彦, 冨山哲男, 吉川弘之
    日本機械学會論文集. C編 社団法人日本機械学会 61 (581) 330 - 337 0387-5024 1995/12 [Refereed][Not invited]
     
    An intelligent CAD system aims at serving a designer as an intelligent, integrated design environment. For building such a system, development of a large scale knowledge base is necessary. To do so, design knowledge must be systematically formalized, made computable, and organized. Design knowledge has two categories ; i.e., design object knowledge and design process knowledge. In our previous reports, we have formalized a computable design process model to represent design process knowledge and the concept of a metamodel dealing with multiple design object models. This paper proposes a fra...
  • Yoshiki Shimomura, Hideaki Takeda, Masaharu Yoshioka, Yasushi Umeda, Tetsuo Tomiyama
    American Society of Mechanical Engineers, Design Engineering Division (Publication) DE 83 (2 Pt 1) 351 - 360 1995 [Refereed][Not invited]
     
    An extension of the Function-Behavior-State (FBS) diagram is proposed for representing a function in the computer aided conceptual design process. The new FBS method uses three additional concepts, namely: function body, function modifier and object entity, to represent the designers' intention more precisely than the original FBS diagram. The performance of the proposed method is complemented by the Functional Evolution Process (FEP) model which represents the different design processes.
  • Masaharu Yoshioka, Masahiko Nakamura, Tetsuo Tomiyama, Hiroyuki Yoshikawa
    American Society of Mechanical Engineers, Design Engineering Division (Publication) DE 53 7 - 14 1993 [Refereed][Not invited]
     
    An intelligent CAD system is not merely a set of intelligent design tools, but rather it must be an intelligent, integrated design environment. This requests it must be equipped with a large scale knowledge base in which design knowledge is intensively and systematically stored. To do so, design knowledge must be systematically formalized, made computable, and organized. The present paper investigates fundamental issues of systematization of design knowledge. Design knowledge has two categories; i.e., design object knowledge and design process knowledge. In our previous reports, we have formalized a computable design process model to represent design process knowledge and the concept of metamodel dealing with multiple design object models. This paper proposes a framework to achieve harmonized combination of these two. In this framework, the computable design process model navigates the metamodel mechanism to generate and modify models of the design object. We describe a prototype system and illustrate an example of ship design.
  • 多重世界を用いた設計過程の論理による形式化
    武田英明, 吉岡真治, 冨山哲男, 吉川弘之
    1991年度人工知能学会(第5回)論文集, 学習院大学、東京, 1991. 1991年6月28日 1991/06/28 [Not refereed][Not invited]
  • 多重世界を用いた設計過程のシミュレーション
    吉岡真治, 河合浩之, 武田英明, 冨山哲男, 吉川弘之
    1991年度精密工学会春季大会講演論文集, 早稲田大学、東京, 1991. 1991年3月26日発表 1991/03/26 [Not refereed][Not invited]

Books etc

Conference Activities & Talks

  • YOSHIOKA Masaharu, MACHI Kojiro, NAGATA Yuuya
    Proceedings of the Annual Conference of JSAI  2023  The Japanese Society for Artificial Intelligence
     
    We have been working on the project to develop a new research paper database that collects papers from the particular domain and extracts important terms to analyze the cooccurrence among them. This system has been applied to the research domain of nanocrystal device development for evaluation. In this paper, to discuss the generality of the proposed system, we have applied the system to the domain of organic chemical reaction development, which is similar to the previous domain. In this domain, there are open databases for representing the chemical substance and information extraction tool. These tools are good for making the initial database, but we confirm it is also necessary to consider the variation of representing the same chemical substance in different string.
  • 知識工学者のための Wikipedia カテゴリの分析とその活用  [Invited]
    吉岡真治
    人工知能学会 第110回人工知能基本問題研究会  2019/09
  • Knowledge exploratory project for nanocrystal device development  [Invited]
    Masaharu YOSHIOKA
    ISMAC 2016 2016 International Symposium on Multimedia and Communication Technology  2016/09
  • Supporting design and engineering by physical concept ontologies  [Invited]
    Masaharu YOSHIOKA
    The Tenth International Symposium on Tools and Methods of Competitive Engineering (TMCE 2014)  2014/05

MISC

  • YOSHIOKA Masaharu, MACHI Kojiro, NAGATA Yuuya  Proceedings of the Annual Conference of JSAI  JSAI2023-  4Xin145  -4Xin145  2023  
    We have been working on the project to develop a new research paper database that collects papers from the particular domain and extracts important terms to analyze the cooccurrence among them. This system has been applied to the research domain of nanocrystal device development for evaluation. In this paper, to discuss the generality of the proposed system, we have applied the system to the domain of organic chemical reaction development, which is similar to the previous domain. In this domain, there are open databases for representing the chemical substance and information extraction tool. These tools are good for making the initial database, but we confirm it is also necessary to consider the variation of representing the same chemical substance in different string.
  • ICTIR  2023
  • Randy Goebel, Yoshinobu Kano, Mi-Young Kim, Juliano Rabelo 0001, Ken Satoh, Masaharu Yoshioka  The Review of Socionetwork Strategies  16-  (1)  109  -110  2022
  • YOSHIOKA Masaharu  JSAI Technical Report, Type 2 SIG  2021-  (55)  5  -5  2021  

    Wikipedia is one of the largest knowledge resources available via the Internet. Several methods have been proposed for extracting knowledge/information from Wikipedia, but methods based on extracting knowledge from the Wikipedia category structure can not utilize whole part of its structure because of the complexity of the relationships between the various categories. In this paper, we briefly review our previous researches on analyzing Wikipedia category and introduce a framework called "Wikipedia Category Ontology" (WCO) that aims to act as a basis for interpreting the Wikipedia category structure. It is based on a classification of category types and relationship types and available online in the form of Linked Open Data at http://wcontology.org/. We also demonstrate the system by using Wikipedia category analysis scenarios.

  • 木村泰知, 渋木英潔, 高丸圭一, 内田ゆず, 乙武北斗, 石下円香, 三田村照子, 吉岡真治, 秋葉友良, 小川泰弘, 佐々木稔, 横手健一, 神門典子, 森辰則, 荒木健治, 関根聡  言語処理学会年次大会発表論文集(Web)  27th-  2021
  • 小板橋佳晃, 吉岡真治, 吉岡真治  言語処理学会年次大会発表論文集(Web)  27th-  2021
  • 高橋翔哉, 吉岡真治, 吉岡真治  人工知能学会人工知能基本問題研究会資料  112-  42  -47  2020/03/08  [Not refereed][Not invited]
  • 中野裕太, 吉岡真治, 吉岡真治, 吉岡真治, 吉岡真治  人工知能学会人工知能基本問題研究会資料  111-  23  -28  2020/01/29  [Not refereed][Not invited]
  • 木村泰知, 渋木英潔, 高丸圭一, 秋葉友良, 石下円香, 内田ゆず, 小川泰弘, 乙武北斗, 佐々木稔, 三田村照子, 横手健一, 吉岡真治, 神門典子, 神門典子  言語処理学会年次大会発表論文集(Web)  26th-  2020
  • 小板橋佳晃, 吉岡真治, 吉岡真治  言語処理学会年次大会発表論文集(Web)  26th-  2020
  • 立浪紀彦, 吉岡真治, 神門典子, ALLAN James, 伊藤正彦  言語処理学会年次大会発表論文集(Web)  26th-  2020
  • 町光二郎, 吉岡真治, 吉岡真治, 吉岡真治, 吉岡真治  言語処理学会年次大会発表論文集(Web)  26th-  2020
  • 町光二郎, 吉岡真治  情報処理学会研究報告(Web)  2020-  (IFAT-139)  2020
  • NAKAGAWA Takanori, KOITABASHI Yoshiaki, YOSHIOKA Masaharu  Proceedings of the Annual Conference of JSAI  2020-  (0)  4H3GS405  -4H3GS405  2020  [Not refereed][Not invited]
     

    Since a category of Wikipedia contains information useful for classifying a Wikipedia pages. A classification method using this category information has been proposed.However, since various types of information are mixed in the Wikipedia category, there is a problem that a category that is not appropriate to be assigned at the same time is added.In this research, we classify more appropriate classes for pages with multiple categories using Wikipedia category ontology, which defines the types of Wikipedia categories and their parent-child relationships. By using this method, we rank the adequacy of the page class belonging to a plurality of categories. To confirm effectiveness of this method, we compar with the classification of SHINRA : Wikipedia Structuring Project.

  • TAKAMARU Keiichi, KIMURA Yasutomo, UCHIDA Yuzu, SASAKI Minoru, YOSHIOKA Masaharu, AKIBA Tomoyoshi, SHIBUKI Hideyuki  Proceedings of the Annual Conference of JSAI  2020-  (0)  4Q3GS901  -4Q3GS901  2020  [Not refereed][Not invited]
     

    In the NTCIR-15 QA Lab-PoliInfo-2, we are planning a Stance Classification Task that estimates the pros and cons of each bill of the group in the Tokyo Metropolitan Assembly. This paper describes the outline of the task and the construction of the data set for the task. Then, this paper analyzes the existence of remarks and their expression forms for expressing approval or disapproval of bills included in the minutes of the Tokyo Metropolitan Assembly in 2018, which is part of the data used in the task.

  • FURUKAWA Kei, YOSHIOKA Masaharu  Proceedings of the Annual Conference of JSAI  2020-  (0)  3Rin470  -3Rin470  2020  [Not refereed][Not invited]
     

    Lots of technical documents have been created and accumulated for the purpose of standardizing and inheriting technologies in many companies across industries. Those documents have certain formats and can be searched by documents retrieval systems. However, since many of those documents are written in specific fields, it's difficult for younger engineers to find the documents related to their tasks if they don't know appropriate key words whereas it's easy for experienced engineers who know the existence of such documents. In this study, we conduct feasibility studies to reveal the above problems with documents in a construction company as an example and introduce the basic idea of the new technical documents retrieval system utilizing construction steps and conditions.

  • KATAYOSE Ryosuke, MASAHARU Yoshioka  Proceedings of the Annual Conference of JSAI  2020-  (0)  3Rin412  -3Rin412  2020  [Not refereed][Not invited]
     

    Stock price forecasting is performed in various forms, such as technical analysis and fundamental analysis using time series data and financial indicators, and text analysis using news information. We propose a method for predicting stock prices by machine learning using multiple technical indicators, taking into account the increasing number of algorithmic trades using technical analysis. At this time, stocks with good and bad fits by technical analysis are classified and evaluated. We consider how to combine multiple technical indicators.

  • 吉岡真治, 尹磊, 原真二郎, 鈴木晃, 高山英紀, 石井真史  言語処理学会年次大会発表論文集(Web)  25th-  2019
  • 狩野芳伸, 吉岡真治, KIM Mi-Young, LU Yao, RABELO Juliano, 清田直希, GOEBEL Randy, 佐藤健  言語処理学会年次大会発表論文集(Web)  25th-  2019
  • 中川嵩教, 小板橋佳晃, 吉岡真治, 吉岡真治  言語処理学会年次大会発表論文集(Web)  25th-  2019
  • 吉岡真治, 吉岡真治, 大久保好章, 尹磊, 原真二郎, 鈴木晃, 高山英紀, 石井真史  応用物理学会秋季学術講演会講演予稿集(CD-ROM)  80th-  2019
  • 吉岡真治, 吉岡真治  人工知能学会人工知能基本問題研究会資料  110th-  2019
  • 吉岡 真治  知能と情報  31-  (2)  23  -23  2019  [Not refereed][Invited]
  • NAKAGAWA Takanori, KOITABASHI Yoshiaki, YOSHIOKA Masaharu  Proceedings of the Annual Conference of JSAI  2019-  (0)  1K2J405  -1K2J405  2019  [Not refereed][Not invited]
     
    <p>Wikipediaカテゴリを分析する研究から、カテゴリには概念分類を表すものだけでなく、人名や組織名といったインスタンスを表すものやその組み合わせが存在しているために、知識工学的観点から利用するためには、そのカテゴリ間の関係について検討することが必要であることが提案された。また、この考え方に基づき、カテゴリの種類、カテゴリの親子関係の種類を整理したデータの構築を行った。本稿では、この構築したデータに基づき、特定の親子関係のみを利用して(あるいは、排除して)できあがるカテゴリ階層構造の性質を議論することで、これまでに提案してきたWikipediaカテゴリに関する分類の枠組の有用性について議論する。</p>
  • 立浪紀彦, 吉岡真治  情報処理学会研究報告(Web)  2018-  (ICS-192)  Vol.2018‐ICS‐192,No.6,1‐4 (WEB ONLY)  2018/06/29  [Not refereed][Not invited]
  • 吉岡真治, 中川嵩教  言語処理学会年次大会発表論文集(Web)  24th-  ROMBUNNO.P12‐1 (WEB ONLY)  2018  [Not refereed][Not invited]
  • NAKAGAWA Takanori, YOSHIOKA Masaharu  Proceedings of the Annual Conference of JSAI  2018-  (0)  2F402  -2F402  2018  [Not refereed][Not invited]
     

    Since Wikipedia is an on-line encyclopedia that covers varieties topics, there are several attempts to extracting knowledge from its contents. Wikipedia category is one type of information that provides topical index for the articles and organized as hierarchical manner. However, since hierarchical structure of Wikipedia category have different characteristics compared to the one for knowledge representation, it may not fully and correctly utilized at this moment. In this paper, we propose a method to extract hierarchical structure that have similar characteristics with knowledge representation.

  • 小野寺大輝, 吉岡真治  言語処理学会年次大会発表論文集(Web)  23rd-  ROMBUNNO.C3‐4 (WEB ONLY)  2017  [Not refereed][Not invited]
  • 吉岡 真治  人工知能学会全国大会論文集  2017-  (0)  1E1OS24a3  -1E1OS24a3  2017  [Not refereed][Not invited]
     

    本試論では、自動運転の車のようなオープンな環境で利用される人工知能を搭載された人工物の製造物責任について検討する。オープンな環境における責任を議論するためには、様々な事態と付随するリスクを認識し、製造者と利用者が適切に分担する枠組が必要になると考えている。本論では、トロッコ問題を例にとり、製造者と利用者がリスクを適切に認識するための枠組とその実現のための人工知能の果たすべき役割について考察する。

  • Thaer M. Dieb, Masaharu Yoshioka, Shinjiro Hara  情報処理学会論文誌  57-  (4)  2016/04/15  
    The process of nanocrystal device development is not well systematized. To support this process, analysis of the information produced by developmental experiments is required. In this study, we constructed an annotated corpus to support the extraction of experimental information from relevant publications. We designed the corpus-construction guidelines by cooperating with a domain expert. We evaluated these guidelines through corpus-construction experiments with graduate students from this domain, and then evaluated the corpus with the domain expert. In the corpus construction experiments, we achieved a sufficient level of Inter-Annotator Agreement by using a loose agreement measure that ignored the term-boundary mismatch problem, and made an agreement corpus that excluded annotations based on misunderstanding the guidelines. The domain expert evaluated this agreement corpus and modified the guidelines based on real examples. Using these guidelines, we finalized the corpus called "NaDev" (Nanocrystal Device development corpus). The NaDev corpus and its construction guidelines will be released via our website, http://nanoinfo.ist.hokudai.ac.jp/. The NaDev corpus aims to support automatic information extraction from publications relevant to nanocrystal device development. This information can be used to solve problems in the nanotechnology domain using the massive availability of fresh information. To the best of our knowledge, this is the first corpus constructed for the development of nanocrystal devices.\n------------------------------This is a preprint of an article intended for publication Journal ofInformation Processing(JIP). This preprint should not be cited. Thisarticle should be cited as: Journal of Information Processing Vol.24(2016) No.3 (online)------------------------------The process of nanocrystal device development is not well systematized. To support this process, analysis of the information produced by developmental experiments is required. In this study, we constructed an annotated corpus to support the extraction of experimental information from relevant publications. We designed the corpus-construction guidelines by cooperating with a domain expert. We evaluated these guidelines through corpus-construction experiments with graduate students from this domain, and then evaluated the corpus with the domain expert. In the corpus construction experiments, we achieved a sufficient level of Inter-Annotator Agreement by using a loose agreement measure that ignored the term-boundary mismatch problem, and made an agreement corpus that excluded annotations based on misunderstanding the guidelines. The domain expert evaluated this agreement corpus and modified the guidelines based on real examples. Using these guidelines, we finalized the corpus called "NaDev" (Nanocrystal Device development corpus). The NaDev corpus and its construction guidelines will be released via our website, http://nanoinfo.ist.hokudai.ac.jp/. The NaDev corpus aims to support automatic information extraction from publications relevant to nanocrystal device development. This information can be used to solve problems in the nanotechnology domain using the massive availability of fresh information. To the best of our knowledge, this is the first corpus constructed for the development of nanocrystal devices.\n------------------------------This is a preprint of an article intended for publication Journal ofInformation Processing(JIP). This preprint should not be cited. Thisarticle should be cited as: Journal of Information Processing Vol.24(2016) No.3 (online)------------------------------
  • 小野寺大輝, 吉岡真治  言語処理学会年次大会発表論文集(Web)  22nd-  P13‐5 (WEB ONLY)  2016  [Not refereed][Not invited]
  • 小野寺大輝, HUANG Le, 吉岡真治  人工知能学会全国大会論文集(CD-ROM)  2016-  (0)  4B45  -4B45  2016  [Not refereed][Not invited]
  • ZHU Tao, DIEB Thaer M, 吉岡真治, 原真二郎  人工知能学会全国大会論文集(CD-ROM)  30th-  ROMBUNNO.1J2‐4  2016  [Not refereed][Not invited]
  • 朱 濤, Dieb Thaer Mostafa, 吉岡 真治, 原 真二郎  人工知能学会全国大会論文集  2016-  (0)  1J24  -1J24  2016  [Not refereed][Not invited]
  • 吉岡 真治, Loban Rhett  人工知能学会全国大会論文集  2015-  (0)  1G42  -1G42  2015  [Not refereed][Not invited]
     

    Web上の百科事典であるWikipediaは、様々な事象に関する知識源であるが、その情報の品質については、編集者に依存する。本研究では、DBPediaにより抽出された構造化情報を用いて、Wikipediaのカテゴリ構造の一貫性を分析する方法を提案するとともに、その分析事例について紹介する。

  • Multifacet comparative analysis of newspaper articles from different conutries - Analysis based on Global Database of Events, Language and Tone (GDELT) - （Awarded JSAI Incentive Award)
    Masaharu Yoshioka, Noriko Kando  The Proceedings of the 8th Symposium on JSAI Interactive Visualization and Mining  SIG-AM-08-0-  24  -31  2014/11  [Not refereed][Not invited]
  • 岸桂太, 吉岡真治  研究報告情報基礎とアクセス技術（IFAT）  2014-  (9)  1  -5  2014/03/22  [Not refereed][Not invited]
     
    多くの特許では、既存の製品の機能向上や好ましくない点の抑制といった効果を目指している。これらの効果に関する記述は、定型的に表現されることが多く、これを特長表現と呼ぶ。本研究では、特長表現から対象と観点を抽出し、それらを用いた特許マップの生成手法を提案する。
  • 吉岡 真治  人工知能学会全国大会論文集  2014-  (0)  2J34  -2J34  2014  [Not refereed][Not invited]
     

    Wikipediaのカテゴリーは、閲覧者にとって有用な分類を提供するものであり、そのカテゴリーは束状の階層関係として表現される。 この関係には、概念間の階層関係を含む様々な関係が存在する。 本発表では、このカテゴリー階層関係の特徴的なパターンを分析すると共に、その分析結果に基づいた日本語Wikipediaオントロジーに関する考察を行う。

  • 岸 桂太, 吉岡 真治  電子情報通信学会技術研究報告 = IEICE technical report : 信学技報  113-  (150)  55  -58  2013/07/22  [Not refereed][Not invited]
     
    多くの特許では、既存の製品の機能向上や好ましくない点の抑制といった効果を目指している。これらの効果に関する記述は、定型的に表現されることが多く、これを特長表現と呼ぶ。本研究では、特長表現を網羅的に集めるための手法を提案するとともに、効果とその対象に注目した特許マップの生成手法を提案する。
  • ナノ知識探索プロジェクト：実験記録からの知識発見 (第3報) – Nanodevice Research Papers Clustering Based on Automatic Paper Annotation –
    Thaer Moustafa Dieb, 吉岡 真治, 原 真二郎  2013年度(第27回)人工知能学会全国大会 (JSAI 2013), 富山, 2013年6月4日-7日, 発表論文集, 1C3-4 (4pp.)  1C3-4-1  -1C3-4-4  2013/06  [Not refereed][Not invited]
  • Daichi KOIKE, KensakuMAKITA, TakehitoUTSURO, Yasuhide KAWADA, Masaharu YOSHIOKA, Noriko KANDO  Proceedings of the 5th Forum for Data Engineering and Information Management  2013/03  [Not refereed][Not invited]
  • ニュース・ツイッター間の対応を考慮したバースト・トピックの同定
    高橋 佑介, 胡 碩, 宇津呂 武仁, 吉岡 真治, 神門 典子  第5回データ工学と情報マネジメントに関するフォーラム---DEIM フォーラム--- 論文集  2013/03  [Not refereed][Not invited]
  • HU Shuo, 高橋佑介, ZHENG LiYi, 宇津呂武仁, 吉岡真治, 神門典子  言語処理学会年次大会発表論文集(Web)  19th-  WEB ONLY C3-1  2013  [Not refereed][Not invited]
  • ZHENG Liyi, HU Shuo, 小池大地, 宇津呂武仁, 吉岡真治, 神門典子  言語処理学会年次大会発表論文集(Web)  19th-  WEB ONLY C3-3  2013  [Not refereed][Not invited]
  • 小池大地, 宇津呂武仁, 河田容英, 吉岡真治  人工知能学会全国大会論文集(CD-ROM)  2013-  (0)  2E14in  -2E14in  2013  [Not refereed][Not invited]
     

    本論文では，検索エンジンAPIを用いたウェブページ収集タスクにおいて，ト ピックモデルを用いたトピックの分布の観点から，できる限り多様なウェブペー ジを収集する方式を提案する．提案方式によって，ニュース・ブログにおける トピックとの関連が強いウェブページを収集することにより，より多様性に富 んだトピックの分布が観測できることを示す．

  • Thaer Mostafa Dieb, 吉岡 真治, 原 真二郎  人工知能学会全国大会論文集  2013-  (0)  1C34  -1C34  2013  [Not refereed][Not invited]
     

    We have been working on the project "Exploring knowledge for nanodevice development" and proposed a framework to annotate useful information (e.g., source material, evaluation parameter, and so on) for analyzing nanocrystal device development papers. In this paper, we conduct nanocrystal device research papers clustering experiment once on annotated papers using automatic annotation results, and once as non-annotated using bag of word approach, and then we compare the results discussing the usefulness of the automatic annotation.

  • 吉岡真治, 神門典子  情報アクセスシンポジウム2012  2012-  (2012)  1  -7  2012/11/30  [Not refereed][Not invited]
  • 吉岡真治, THAER Moustafa Dieb, 原真二郎  Designシンポジウム講演論文集  2012-  87  -91  2012/10/16  [Not refereed][Not invited]
  • ナノ知識探索プロジェクト：ナノデバイス開発記録からの知識発見
    吉岡 真治, Thaer Moustafa Dieb, 原 真二郎  Designシンポジウム2012, 京都, 2012年10月16日-17日, 発表論文集 (5pp.)  1  -5  2012/10  [Not refereed][Not invited]
  • 吉岡 真治, 劉 亦奇, 神門 典子  情報処理学会論文誌データベース（TOD）  5-  (3)  141  -148  2012/09/28  [Not refereed][Not invited]
     
    Recently, due to the higher demand for geographic information system, it is necessary to have a good geographical database for such systems. GeoNames is one of the largest geographical database as Open Data. This database is also used for constructing web ontology by adding links to the Wikipedia page as a part of Linked Open Data. Even though, here are several attempts to find links automatically, the number of links between GeoNames and Wikipedia is not sufficient. In this paper, we propose an automatic link discovery method to use Wikipedia categories to identify the correspondence between Wikipedia page and GeoNames entry. We also propose to use this method for inappropriate link detection. Link data obtained in this paper is already sent to the manager of GeoNames and a part of the result is used for updating the site.
  • HU Shuo, ZHENG Liyi, TAKAHASHI Yusuke, KOIKE Daichi, MAKITA Kensaku, UTSURO Takehito, YOSHIOKA Masaharu  IEICE technical report. Natural language understanding and models of communication  112-  (196)  25  -30  2012/08/23  [Not refereed][Not invited]
     
    This paper studies issues regarding topic modeling of information flow in multilingual news and blog streams. If someone wants to find differences in the topics of Japanese and Chinese news / blog, it is usually necessary for him/her to carefully watch every article in Japanese and Chinese news streams at every moment. In such a situation, topic models such as LDA (Latent Dirichlet Allocation) and DTM (dynamic topic model) are quite effective in estimating distribution of topics over a document collection such as articles in a news / blog stream. Based on the results of estimating distribution of topics in Japanese / Chinese news streams, this paper proposes how to analyze cross-lingual alignment of topics in time series Japanese / Chinese news streams. This paper also proposes how to analyze differences of concerns and opinions among various Chinese news sites as well as blog hosts.
  • 吉岡真治  研究報告情報基礎とアクセス技術（IFAT）  2012-  (1)  1  -5  2012/07/25  [Not refereed][Not invited]
     
    現在、Wikipedia の情報を中心に様々な Open Data を関連づけて活用する Linked Open Data の研究が行われている。本発表では、Wikipedia のページが表現する情報の粒度が与える影響について考察を行う。
  • Tetsuya Sakai, Hideo Joho, Noriko Kando, Tsuneaki Kato, Akiko Aizawa, Tomoyoshi Akiba, Isao Goto, Fuminori Kimura, Teruko Mitamura, Hiromitsu Nishizaki, Hideki Shima, Masaharu Yoshioka, Shlomo Geva, Ling-Xiang Tang, Andrew Trotman, Yue Xu  IPSJ SIG Notes  2012-  (5)  1  -15  2012/03/19  [Not refereed][Not invited]
     
    NTCIR is a sesquiannual evaluation workshop series on information access which started in 1999 and is run by National Institute of Informatics (NII). The ninth workshop (NTCIR-9) was concluded in December 2011 and the accepted tasks for the tenth workshop (NTCIR-10) have been announced in January 2012. This paper provides a brief summary of NTCIR-9, and introduces some latest developments from NTCIR-10.
  • DIEB THAER M, YOSHIOKA MASAHARU, HARA SHINJIRO  言語処理学会年次大会発表論文集  18th (CD-ROM)-  ROMBUNNO.B5-2  2012/03/13  [Not refereed][Not invited]
  • HU Shuo, 高橋佑介, 牧田健作, 横本大輔, 宇津呂武仁, 吉岡真治  言語処理学会年次大会発表論文集  18th (CD-ROM)-  ROMBUNNO.F1-5  2012/03/13  [Not refereed][Not invited]
  • 高村慎太郎, 吉岡真治, 関洋平  言語処理学会年次大会発表論文集  18th (CD-ROM)-  ROMBUNNO.P1-11  2012/03/13  [Not refereed][Not invited]
  • 高橋佑介, 横本大輔, 宇津呂武仁, 吉岡真治  言語処理学会年次大会発表論文集  18th (CD-ROM)-  ROMBUNNO.F1-4  2012/03/13  [Not refereed][Not invited]
  • 小池大地, 横本大輔, 牧田健作, 鈴木浩子, 宇津呂武仁, 河田容英, 吉岡真治, 福原知宏  言語処理学会年次大会発表論文集  18th (CD-ROM)-  ROMBUNNO.A5-1  2012/03/13  [Not refereed][Not invited]
  • Idetification of Burst in the Chronological Topic Model
    Yusuke Takahashi, Daisuke Hashimoto, Takehito Utsuro, Masaharu Yoshioka, Yoshihide Kawada, Noriko Kando, Tomohiro Fukuhara, Hiroshi Nakagawa, Yoji Kiyota, Noriko Kando  Proceedings of the 4th Data Engineering and Information Management Forum (DEIM Forum 2012)  F5-5  2012/03  [Not refereed][Not invited]
  • Bird' Eye View Analysis on Topi Distribution of Bloggers
    Kensaku Makita, Daisuke Hashimoto, Hiroko Suzuki, Takehito Utsuro, Noriko Kando, Yoshihide Kawada, Tomohiro Fukuhara, Hiroshi Nakagawa, Masaharu Yoshioka, Yoji Kiyota  Proceedings of the 4th Data Engineering and Information Management Forum (DEIM Forum 2012)  F7-4  2012/03  [Not refereed][Not invited]
  • Topic Correlation and Transitions in News Blogs: With the Exmples Related to the Disasters
    Daichi Koike, Daisuke Hashimoto, Kensaku Makita, Hiroko Suzuki, Takhito Utsuro, Yohihide Kawada, Masaharu Yoshioka, Noriko Kando, Tomohiro Fukuhara, Hiroshi Nakagawa, Yoji Kiyota, Yohei Seki  Proceedings of the 4th Data Engineering and Information Management Forum (DEIM Forum 2012)  F1-4  2012/03  [Not refereed][Not invited]
  • 藤原 嵩大, 吉岡 真治  人工知能学会全国大会論文集  26-  1  -4  2012  [Not refereed][Not invited]
  • 小池大地, 牧田健作, 宇津呂武仁, 河田容英, 吉岡真治, 福原知宏  人工知能学会全国大会論文集(CD−ROM)  2012-  (0)  3E1R69  -3E1R69  2012  [Not refereed][Not invited]
     

    本論文では，震災に関する話題についての時系列のニュース記事集合，および， ブログ記事集合を対象として，トピックモデルを用いたトピック同定を行い， ニュース・ブログの間での話題の相関を分析する．本論文では，特に，各トピッ クと密接に対応するニュース記事集合・ブログ記事集合における話題がどの程度 まとまっているか，という点と，時系列バーストの有無との相関について分析を 行った結果を報告する．

  • 吉岡 真治, 神門 典子, 関 洋平  研究報告 情報基礎とアクセス技術（IFAT）  2011-  (2)  1  -8  2011/07/26  [Not refereed][Not invited]
     
    本研究では、日中韓を中心とした複数国の新聞を比較分析する NSContrast を提案している。本システムでは、実際に毎日の新聞を新聞社のサイトから収集し、その日に注目を浴びている単語のリストや、そのような単語を含む記事数の増減や関連語などを表示することにより、従来のニュースポータルとは異なった観点からの新聞記事分析を可能にするものである。本稿では、本システムの有効性を検証するために行ったユーザ実験とその分析結果について報告する。We have already proposed a system News Site Contrast(NSContrast) that analyzing the characteristic difference between each news site and all others. The system make news article database by retrieving articles from multiple news sites in daily basis. In addition, the system have functions to analyze these articles by showing the hot keywords, related keywords for the retrieved documents. In this paper, we report the result of user experiment and analysis of the results.
  • DIEB THAER M, YOSHIOKA MASAHARU  言語処理学会年次大会発表論文集  17th (CD-ROM)-  ROMBUNNO.P1-29  2011/03  [Not refereed][Not invited]
  • 特定トピックに関するブログ記事集合の観点分類におけるWikipediaの利用
    横本 大輔, 林 東權, 牧田 健作, 宇津呂 武仁, 河田 容英, 福原 知宏, 神門 典子, 吉岡 真治, 中川 裕志, 清田 陽司  第3回データ工学と情報マネジメントに関するフォーラム  2011/02  [Not refereed][Not invited]
  • Daisuke Yokomoto, Dangkwon Lim, Kensaku Makita, Takehito Utsuro, Yasuhide Kawada, Tomohiro Fukuhara, Noriko Kando, Masaharu Yoshioka, Hiroshi Nakagawa, Yoji Kiyota  Proceedings of DEIM Forum 2011  2011/02  [Not refereed][Not invited]
  • 原口誠, 大久保好章, 富田悦次, 吉岡真治  人工知能学会人工知能基本問題研究会資料  84th-  2011
  • 竹中 均, 吉岡 真治  人工知能学会全国大会論文集  2011-  (0)  2J3NFC22  -2J3NFC22  2011  [Not refereed][Not invited]
     

    東京の記述に日本という説明を加えないように、地名を含む情報を記載する場合に、その包含関係の情報を明示的に記述しないことが多い。しかし、これらの地名に関連した情報を検索する場合には、このような暗黙に仮定している関係の情報を利用する必要がある。本研究では、Wikipediaを対象にして、このような検索に有用な地名の包含関係情報の抽出方法を提案する。

  • 吉岡真治, THAER Dieb, 冨岡克広, 原真二郎, 福井孝志  人工知能学会全国大会論文集(CD−ROM)  25-  1  -4  2011  [Not refereed][Not invited]
  • Yusuke Takahashi, Daisuke Yokomoto, Takehito Utsuro, Masaharu Yoshioka  IPSJ SIG Notes  6-  (6)  1  -6  2011  [Not refereed][Not invited]
     
    Among various types of recent information explosion, that in news stream is also a kind of serious problems. This paper studies issues regarding two types of modeling of information flow in news stream, namely, burst analysis and topic modeling. First, when one wants to detect a kind of topics that are paid much more attention than usual, it is usually necessary for him/her to carefully watch every article in news stream at every moment. In such a situation, it is well known in the field of time series analysis that Kleinberg's modeling of bursts is quite effective in detecting burst of key...
  • 吉岡真治, 伊藤正彦, SEBAG Michele, FEKETE Jean‐Daniel  情報処理学会研究報告(CD-ROM)  2009-  (6)  ROMBUNNO.FI-98,8  2010/04/15  [Not refereed][Not invited]
  • 吉岡 真治  全国大会講演論文集  72-  (0)  301  -302  2010/03/08  [Not refereed][Not invited]
  • YOSHIOKA MASAHARU, ITOH MASAHIKO, SEBAG MICHELE, FEKETE JEAN-DANIEL  研究報告情報学基礎（FI）  98-  (8)  H1  -H6  2010/03/04  [Not refereed][Not invited]
     
    多次元尺度構成法(MDS)は、高次元のデータを2次元で表現する手法である。しかし、ユーザが初期の出力結果に満足できない場合が存在する。この問題に対し、本発表では、ユーザにとって分類に役立つ次元を学習する距離尺度学習の手法を応用することによって、MDSの結果をインタラクティブに操作する方法を提案する。また、このシステムを論文の類似度を元にして作成した複数研究者の関係を可視化するシステムに適用した結果について紹介する。Multi Dimensional Scaling (MDS) is one of the popular method for visualizing multidimensional data into 2D world. However, the user may not satisfy the visualization results. In this paper, we propose a novel interactive method for modifying MDS visualization results by using the concept of distance metric learning. We also demonstrate our researcher visualization system based on this algorithm.
  • 吉岡 真治  電子情報通信学会技術研究報告. AI, 人工知能と知識処理  109-  (424)  43  -46  2010/02/15  [Not refereed][Not invited]
     
    辞書を利用する機械翻訳システムにおいて、未登録語が多く存在する文書の翻訳結果は、質の悪いものとなる。しかし、辞書のメンテナンスは非常にコストがかかり、新語や、時事的な単語を辞書に継続的に登録することは容易ではない。本研究では、このような機械翻訳の辞書を補完するという目標のもとに、Wikipediaの言語間リンクを用いて中日の翻訳辞書を作成する方法を提案する。また、この翻訳辞書の活用事例として、多言語ニュースサイトの比較分析システムNSContrastの紹介を行う。
  • 吉岡 真治, 冨岡 克広, 原 真二郎, 福井 孝志  人工知能学会全国大会論文集  2010-  (0)  1B33  -1B33  2010  [Not refereed][Not invited]
     

    ナノ結晶デバイスの開発プロセスにおいては、製造プロセスがデバイスの性質に大きな影響を与えるため、試行錯誤的な実験の繰り返しが必要とされる。しかし、その試行錯誤に関する知識はあまり明示化されていないため、熟練者の「匠の技」である暗黙知に依存している。本研究では、デバイス生成のための実験を行った記録を情報源として、この暗黙知の明示化を支援する実験記録データの活用システムを提案する。

  • NAKASAKI HIROYUKI, ABE YUSUKE, UTSURO TAKEHITO, KAWADA YASUHIDE, FUKUHARA TOMOHIRO, KANDO NORIKO, YOSHIOKA MASAHARU, NAKAGAWA HIROSHI, KIYOTA YOJI  IPSJ SIG Notes  194-  (9)  I1  -I8  2009/11/16  [Not refereed][Not invited]
     
    Among other domains and topics on which some issues are frequently argued in the blogosphere, the domain of crime is one of the most seriously discussed by various kinds of bloggers. Such information on crimes in blogs is especially valuable for outsiders from abroad who are not familiar with cultures and crimes in foreign countries. This paper proposes a multilingual framework of categorizing people's concerns, reports, and experiences on crimes in their own blogs. First, we refer to Wikipedia as a multilingual terminological knowledge base, and search for Wikipedia entries describing crim...
  • SATO YUKI, YOKOMOTO DAISUKE, NAKASAKI HIROYUKI, UTSURO TAKEHITO, YOSHIOKA MASAHARU, FUKUHARA TOMOHIRO, KANDO NORIKO, NAKAGAWA HIROSHI, KIYOTA YOJI  IPSJ SIG Notes  194-  (10)  J1  -J8  2009/11/16  [Not refereed][Not invited]
     
    We study complementary navigation of news and blog, where Wikipedia entries are utilized as fundamental knowledge source for linking news articles and blog feeds/posts. In the proposed framework, given a topic as the title of a Wikipedia entry, its Wikipedia entry body text is analyzed as fundamental knowledge source for the given topic, and terms strongly related to the given topic are extracted. Those terms are then used for ranking news articles and blog posts. In the scenario of complementary navigation from a news article to closely related blog posts, Japanese Wikipedia entries are ra...
  • 角谷和俊, 吉岡真治, 佐藤哲司  情報処理学会論文誌データベース（TOD）  2-  (3)  i  -iii  2009/09/30  [Not refereed][Not invited]
  • 角谷 和俊, 吉岡 真治, 佐藤 哲司  情報処理学会論文誌データベース（TOD）  2-  (2)  i  -iii  2009/06/29  [Not refereed][Not invited]
  • SUMIYA K, YOSHIOKA M, SATO T  情報処理学会論文誌. データベース  2-  (2)  i  -iii  2009/06/29  [Not refereed][Not invited]
  • YOSHIOKA MASAHARU  IPSJ SIG Notes  71-  (1)  A1  -A8  2009/06/05  [Not refereed][Not invited]
     
    In this research we have already proposed a news site analysis system NSContrast that characterizes the differences among news sites from multiple countries. In this analysis system, news articles database are constructed by using following three types of news sites data; Japanese news articles written in Japanese, foreign ones written in Japanese and machine translated version of foreign ones written in foreign language. In this paper, we discuss how to evaluate the effectiveness of the NSContrast and briefly introduced the result of user experiment study.
  • 角谷 和俊, 吉岡 真治, 佐藤 哲司  情報処理学会論文誌データベース（TOD）  2-  (1)  i  -iii  2009/03/31  [Not refereed][Not invited]
  • OHTSUKA Hiroki, YOSHIOKA Masaharu  IPSJ SIG Notes  94-  (35)  9  -14  2009/03/18  [Not refereed][Not invited]
     
    A Link Structure among Web documents represents the relationship among these documents. Web structure analysis methods such as PageRank use this information. However, most of these methods pay little attention to the type of link (e.g., most of the methods use simple classification such as link to the same site or not). In this research, we proposed a method to classification criteria of anchor texts based on the role analysis of the text in Web documents for better link structure analysis and the users' behavior analysis on the Web page. In this paper, we propose an automatic classification method based on the result of manually classification experiment. In order to evaluate this method, we compare the result of automatic classification results with manually classified ones.
  • 吉岡真治  言語処理学会年次大会発表論文集  15th-  494-497  -497  2009/03/02  [Not refereed][Not invited]
  • 佐藤由紀, 中崎寛之, 川場真理子, 宇津呂武仁, 吉岡真治, 福原知宏, 中川裕志, 神門典子  言語処理学会年次大会発表論文集  15th-  300-303  2009/03/02  [Not refereed][Not invited]
  • 吉岡 真治  人工知能学会全国大会論文集  2009-  (0)  2G1NFC58  -2G1NFC58  2009  [Not refereed][Not invited]
     

    本研究では、様々な国から発信される多言語ニュースに対し、機械翻訳システムを利用して対照比較を行うNSContrastの研究を行っている。本システムの分析には、翻訳間違いや表記の違いが解析に大きな影響を与えるため、正規化が必要である。本稿では、Wikipediaの言語間リンクを用いた翻訳用辞書の生成や、redirectの関係を用いた表記のバリエーションへの対応を行い、実運用における評価を行う。

  • Sugiyama Aki, Shimomura Yoshiki, Chiba Ryosuke, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2009-  (0)  211  -212  2009  [Not refereed][Not invited]
     
    創造的設計のためには，異領域知識間の柔軟な統合が重要であると考えられており，著者らはそれを実現する統合計算機環境の開発を行っている．本稿では，知識中に記述される語彙情報から階層構造と単位次元により表現する視点情報を計算機により自動的に推定・付与するシステムを提案・開発する．また，視点情報に基づき生成する仮説的知識の創造性と有用性を可変とする操作性を，事例を用いた検証により確認する．
  • 吉岡真治  人工知能学会全国大会論文集(CD−ROM)  23-  1  -4  2009  [Not refereed][Not invited]
  • Takahashi Kiyotaka, Sugiyama Aki, Shimomura Yoshiki, Tateyama Takeshi, Chiba Ryosuke, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2009-  (0)  209  -210  2009  [Not refereed][Not invited]
     
    前報までに、著者らは創造的設計の総合支援環境であるUniversal Abduction Studio（UAS）で利用するための設計に有用な情報を、Web上から自動取得し、当該システム用の知識ベースを構築するための基本手法を提案した。本稿では、上記提案手法を計算機上の一連のソフトウェアとして実装した内容と、実装システムにより取得された知識の設計上の利用可能性のさらに詳細な検証結果について報告する。
  • 角谷和俊, 吉岡真治, 佐藤哲司  情報処理学会論文誌データベース（TOD）  1-  (3)  i  -iii  2008/12/26
  • 角谷和俊, 吉岡真治, 佐藤哲司  情報処理学会論文誌データベース（TOD）  1-  (2)  i  -iii  2008/09/30  [Not refereed][Not invited]
  • 原口 誠, 吉岡 真治  人工知能基本問題研究会  70-  (0)  49  -52  2008/07/04  [Not refereed][Not invited]
  • 角谷和俊, 吉岡真治, 佐藤哲司  情報処理学会論文誌データベース（TOD）  1-  (1)  i  -iii  2008/06/26  [Not refereed][Not invited]
  • 井川 拓, 原口 誠, 吉岡 真治  ことば工学研究会  28-  (0)  31  -34  2008/03/28  [Not refereed][Not invited]
  • 吉岡真治  言語処理学会年次大会発表論文集  14th-  592-595  2008/03/17  [Not refereed][Not invited]
  • SUMIYA K., YOSHIOKA M., SATO T.  情報処理学会論文誌データベース（TOD）  49-  (7)  i  -iii  2008/03/15  [Not refereed][Not invited]
  • AKIBA RYUJI, HARAGUCHI MAKOTO, YOSHIOKA MASAHARU  人工知能学会人工知能基本問題研究会資料  69th-  39-42  2008/03/14  [Not refereed][Not invited]
  • OHTSUKA Hiroki, YOSHIOKA Masaharu  全国大会講演論文集  70-  (0)  209  -210  2008/03/13  [Not refereed][Not invited]
  • 杉山明紀, 下村芳樹, 吉岡真治, 武田英明  日本機械学会設計工学・システム部門講演会論文集(CD-ROM)  18th-  2008
  • Masaharu Yoshioka  INFORMATION RETRIEVAL TECHNOLOGY  4993-  508  -513  2008  [Refereed][Not invited]
     
    In order to utilize news articles from multiple news sites, it is better to understand the characteristics of each news site. In this paper, a concept of contrast set mining is applied for analyzing the characteristic difference between each news site and all others. The News Site Contrast (NSContrast) system is also proposed based on this mining technique. This system is applied to a news article database constructed from multiple news sites from different countries in order to evaluate its effectiveness.
  • 大塚 博紀, 吉岡 真治  人工知能学会全国大会論文集  8-  (0)  203  -203  2008  [Not refereed][Not invited]
     
    本研究では、Webのリンク構造を分類して利用することを目標として、アンカーテキストの情報を用いたリンクの役割分析を行っている。本発表では、役割分類の基準とそれに基づいた自動分類の手法について提案する。
  • 杉山明紀, 下村芳樹, 千葉龍介, 吉岡真治, 武田英明  Designシンポジウム講演論文集  2008-  441-446  2008  [Not refereed][Not invited]
  • OTSUKA HIROKI, YOSHIOKA MASAHARU  人工知能学会全国大会論文集(CD−ROM)  22-  1  -3  2008  [Not refereed][Not invited]
  • Nakanishi Masatoshi, Suzuki Ryo, Simomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2008-  (0)  435  -436  2008  [Not refereed][Not invited]
     
    本研究では創造的設計を総合的に支援する計算機環境の構築を目指している。著者らは、これまでに創造的設計を支援するための知識記述形式や推論機構、オントロジー利用の枠組みなどを提案してきた。しかしながらこれらを用いて創造的な設計を支援するためには以上の構成要素群を統合する必要がある。本稿では創造的設計を支援するための統合計算機環境を提案し、その具体的な実装構造を例示する。
  • Takahashi Kiyotaka, Suzuki Ryou, Shimomura Yoshiki, Tateyama Takeshi, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2008-  (0)  437  -438  2008  [Not refereed][Not invited]
     
    In recent years, comprehensive problem solving by a designer has been required as the demand for an artifact grows and complicates more. In the above background, we have proposed Universal Abduction Studio (UAS), a computer environment, which supports a creative design synthetically. However, it is difficult for a designer to acquire the multiple domain knowledge on the grounds that the knowledge is accumulated manually. This paper proposes a method of expanding the design knowledge database used in UAS by extracting useful information for design from the resource on the web. Also, the authors propose a filtering method for reducing the information which isn't useful for design from the database. At the end, the effectively of the methods is verified.
  • Sugiyama Aki, Shimomura Yoshiki, Chiba Ryosuke, Takeda Hideaki, Yoshioka Masaharu  Proceedings of JSPE Semestrial Meeting  2008-  (0)  647  -648  2008  [Not refereed][Not invited]
     
    著者らは，創造的設計を支援する統合計算機環境の構築を行っている．これまでに，語彙情報をタグ付けした知識から概念類似度を基に生成される仮説的知識の有用性向上手法を提案しているが，場面に即した仮説的知識を効率的に取得するためには語彙の属性に基づくタグ付け手法とそれを用いた推論が有効である．本稿では，知識に対して語彙情報に基づく物理量をタグ付けする知識表現手法を提案し，その有効性を検証する．
  • Takahashi Kiyotaka, Sugiyama Aki, Shimomura Yoshiki, Tateyama Takeshi, Chiba Ryosuke, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2008-  (0)  649  -650  2008  [Not refereed][Not invited]
     
    近年，人工物に対する要求の高度化・複雑化に伴い，設計者に対する負担が一層増大している．以上の背景のもと，著者らは創造的設計を総合的に支援する計算機環境Universal Abduction Studio（UAS）を提案している．本稿では，設計に有用な情報をWeb情報より自動的に取得，UASで利用可能な知識表現に変換することにより，UASの設計知識データベースを効率的に拡充するための手法を提案する．
  • SUGIYAMA Aki, SHIMOMURA Yoshiki, YOSHIOKA Masaharu, TAKEDA Hideaki  The Proceedings of Design & Systems Conference  2008-  (0)  181  -185  2008  [Not refereed][Not invited]
     
    In order to realize a creative design, it is thought to be important to unify the knowledge from different knowledge areas flexibly. The computer environment that supports designers in an integrated way has been developed. The extended method of the knowledge representation forms for increasing the usefulness of the hypothetical knowledge from the knowledge tagged with vocabulary information was proposed. However, the current method isn't always capable of obtaining the useful knowledge, since it doesn't consider the context. In order to acquire the hypothetical knowledge adapted to the context efficiently, the tagging technique based on the attribute of a vocabulary and the reasoning using it are effective. In this paper, the authors propose knowledge representation forms for tagging physical quantity based on vocabulary describing knowledge, and verify the usefulness of the proposed method.
  • TAKAHASHI Kiyotaka, SHIMOMURA Yoshiki, TATEYAMA Takeshi, YOSHIOKA Masaharu, TAKEDA Hideaki  The Proceedings of Design & Systems Conference  2008-  (0)  305  -309  2008  [Not refereed][Not invited]
     
    In recent years, comprehensive problem solving by a designer has been required as the demand for an artifact grows and complicates more. In the above background, we have proposed Universal Abduction Studio (UAS), a computer environment, which supports a creative design synthetically. However, it is difficult for designers to acquire knowledge in multiple domains efficiently because they have to accumulate the design knowledge by the handwork. Thus far, we proposed Web-based design knowledge information database extension method. In this paper, we propose the conversion method acquired design knowledge information from the resource on Web into knowledge expression for UAS.
  • SUMIYA K., YOSHIOKA M., SATO T.  情報処理学会論文誌データベース（TOD）  48-  (20)  i  -iii  2007/12/15  [Not refereed][Not invited]
  • 吉岡 真治  Human interface  9-  (4)  269  -274  2007/11/25  [Not refereed][Not invited]
  • SUGIYAMA AKI, SUZUKI RYO, NAKANISHI MASATOSHI, SHIMOMURA YOSHIKI, YOSHIOKA MASAHARU, TAKEDA HIDEAKI  日本機械学会設計工学・システム部門講演会講演論文集  17th-  286-287  2007/10/30  [Not refereed][Not invited]
  • 杉山明紀, 鈴木遼, 中西雅俊, 下村芳樹, 吉岡真治, 武田英明  精密工学会大会学術講演会講演論文集  2007-  G67  2007/09/03  [Not refereed][Not invited]
  • YOSHIOKA Masaharu  IPSJ Magazine  48-  (8)  920  -921  2007/08/15  [Not refereed][Not invited]
  • 適合的汎化に基づく情報検索システムの実験的研究
    吉岡 真治, 原口誠  NICT-EDR電子化辞書シンポジウム  --  (-)  67  -76  2007/04  [Not refereed][Not invited]
  • 吉岡真治  言語処理学会第13回年次大会ワークショップ (W2)  27  -30  2007/03  [Not refereed][Not invited]
  • Masaharu Yoshioka, Ikuko Yairi  Journal of Japanese Society for Artificial Intelligence  22-  (2)  208  -208  2007/03/01  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, YAIRI Ikuko  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  22-  (2)  208  -208  2007/03/01  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, YAIRI Ikuko  Journal of the Japanese Society for Artificial Intelligence  22-  (2)  208  -208  2007  [Not refereed][Not invited]
  • 吉岡 真治  人工知能  22-  (1)  15  -15  2007  [Not refereed][Not invited]
  • 吉岡 真治, Masaharu Yoshioka  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  22-  (1)  15  -15  2007/01/01  [Not refereed][Not invited]
  • SUGIYAMA Aki, SUZUKI Ryo, NAKANISHI Masatoshi, SHIMOMURA Yoshiki, YOSHIOKA Masaharu, TAKEDA Hideaki  The Proceedings of Design & Systems Conference  2007-  (0)  286  -287  2007  [Not refereed][Not invited]
     
    In order to realize a creative design, it is thought to be important to unify the different knowledge group flexibly, and the concept of the computer environment which supports designers in an integrated way has been proposed. Our project, called Universal Abduction Studio, aims at assisting designers in obtaining design solutions through the dynamic interaction between a system and a designer. In this report, we propose the extended method of the knowledge representation forms for increasing the usefulness of hypothetical knowledge, after considering the existing reasoning environment. Finally, the usefulness of the proposed method is verified.
  • Sugiyama Aki, Suzuki Ryo, Nakanisi Masatosi, Shimomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2007-  (0)  523  -524  2007  [Not refereed][Not invited]
     
    創造的な設計を実現するためには異領域知識間の柔軟な統合が重要と考えられ，それを総合的に支援する計算機環境であるUniversal Abduction Studioの概念が提案されている．本稿では，これまでに著者らが提案した推論機構に関する考察を行なった上で，計算機が導出する仮説的知識の有用性を，各分類と知識テンプレートの導入により向上させる知識表現形式の拡張方法を提案し，その有効性を検証する．
  • Takahashi Kiyotaka, Suzuki Ryou, Shimomura Yoshiki, Tateyama Takeshi, Yoshioka Masaharu, Takeda Hideaki  Proceedings of the Conference of Transdisciplinary Federation of Science and Technology  2007-  (0)  45  -45  2007  [Not refereed][Not invited]
     
    In recent years, comprehensive problem solving by a designer has been required as the demand to an artifact develops and complicates more. In the above background, we have proposed Universal Abduction Studio (UAS), a computer environment that supports a creative design synthetically. This paper proposes the method of expanding the design knowledge database used in UAS by extracting useful information for design from the resource on the web.
  • 吉岡真治, 落合翔平  精密工学会大会学術講演会講演論文集  2007-  (0)  513  -514  2007  [Not refereed][Not invited]
     
    TRIZなどに代表されるようにアイディア検索のために特許を利用する方法が提案されている。このような状況では、表面上の表記に注目した同分野の特許を検索するのではなく、類似の問題を扱っている特許を検索することが重要となる。本研究では、アイディアの根幹にある物理法則をキーとした検索を考え、学習により文書に対して関連物理法則のラベルをつけ、分類する方法を提案する。
  • Yoshioka Masaharu  IPSJ SIG Notes  2006-  (94)  27  -33  2006/09/12  [Not refereed][Not invited]
     
    Most significant characteristics of the Web information is description about a relationship among different pages by using hyperlinks and anchor texts. In this research, anchor texts extracts from large web test collection NTCIR nw100g are investigated for a proposal of a functional classification. A strategy for automatic classification of anchor texts in html pages and systems that utilize this classification are also proposed in this paper.
  • YOSHIOKA Masaharu  IPSJ SIG Notes  84-  (94)  27  -33  2006/09/12  [Not refereed][Not invited]
     
    Most significant characteristics of the Web information is description about a relationship among different pages by using hyperlinks and anchor texts. In this research, anchor texts extracts from large web test collection NTCIR nw100g are investigated for a proposal of a functional classification. A strategy for automatic classification of anchor texts in html pages and systems that utilize this classification are also proposed in this paper.
  • 吉岡真治, 佐藤孝彦, 森本憲悟, 武田英明, 下村芳樹  人工知能学会人工知能基本問題研究会資料  62-  (0)  71  -76  2006/03/27  [Not refereed][Not invited]
  • 伊東真樹, 森本憲悟, 下村芳樹, 武田英明, 吉岡真治  精密工学会大会学術講演会講演論文集  2006-  M22  2006/03/01  [Not refereed][Not invited]
  • 中西雅俊, 森本憲悟, 下村芳樹, 武田英明, 吉岡真治  精密工学会大会学術講演会講演論文集  2006-  M21  2006/03/01  [Not refereed][Not invited]
  • 森本憲悟, 下村芳樹, 吉岡真治, 武田英明, 上田完次  精密工学会大会学術講演会講演論文集  2006-  M23  2006/03/01  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, SATOH Takahiko, SUGIYAMA Aki, MORIMOTO Kengo, TAKEDA Hideaki, SHIMOMURA Yoshiki  Designシンポジウム講演論文集  2006-  (0)  95  -100  2006  [Not refereed][Not invited]
     
    In order to design new artifacts that are different from existing ones, it is crucial to have new knowledge for the design problem. We have already confirmed that combination of existing knowledge from different domain may useful for this process by analyzing design cases. In this paper, we briefly review our hypothetical knowledge generation support environment "Universal Abduction Studio" (UAS) and propose an automatic method to organize design document for this environment.
  • NAKANISHI Masatoshi, SUZUKI Ryo, SHIMOMURA Yoshiki, TAKEDA Hideaki, YOSHIOKA Masaharu  The Proceedings of Design & Systems Conference  2006-  (0)  206  -207  2006  [Not refereed][Not invited]
     
    We argue that abduction for integrating theories can be a basic principle to formalize creative design processes. Based on this principle, we have proposed Universal Abduction Studio, a design environment in which designers combine different theories to arrive at better design. In this new approach to computational support of conceptual design, the system should offer various types of abductive reasoning from which designers can select an interesting design method. This paper aims at proposing the method of handling terminology by general common ontology and managing in the multiple. We also discuss technologies to implement proposed method of managing terminology.
  • 吉岡 真治, 佐藤 孝彦, 森本 憲悟, 武田 英明, 下村 芳樹  人工知能学会全国大会論文集  6-  (0)  202  -202  2006  [Not refereed][Not invited]
     
    従来の設計解とは本質的に異なる創造的な設計解を作るためには、既存の知識の単純な組合わせではなく、新たな設計知識の利用が不可欠である。本研究では、これまで、異なる領域知識間に存在する類似性をキーとしたアナロジーを用いて、仮説的設計知識の生成手法を提案してきた。本発表では、このアナロジーを行なうために必要な文書情報のタグ付けを自動化するための方法論について述べる。
  • Itou Masaki, Morimoto Kengo, Shimomura Yoshiki, Takeda Hideaki, Yoshioka Masaharu  Proceedings of JSPE Semestrial Meeting  2006-  (0)  1001  -1002  2006  [Not refereed][Not invited]
     
    設計課題に関連する領域知識を用いるだけでは,創造的な設計解を得ることは困難である.本研究では創造的設計実現のためには,設計課題に関する領域知識とそれとは異なる領域知識の統合が重要であると考え,これを総合的に支援する計算機環境の構築を目指している.本稿ではより高度な知識統合を実現するために,知識の意味的構造に基づく統合を可能とする知識表現形式とその形式に従い知識を記述するためのエディタを提案する.
  • Morimoto Kengo, Shimomura Yoshiki, Takeda Hideaki, Yoshioka Masaharu, Ueda Kanji  Proceedings of JSPE Semestrial Meeting  2006-  (0)  1003  -1004  2006  [Not refereed][Not invited]
     
    It is important to unify different knowledge groups, which exist independently, and to extend knowledge gradually in order to obtain creative design solutions that are not obtained only with the existing design knowledge. With the aim of development of ″Universal Abduction Studio″, which is a system supporting creative design, we propose a method for creating hypotheses through interactions between designer and machine. We develop a prototype system and demonstrate that the system can create new ideas.
  • YOSHIOKA MASAHARU, SATO TAKAHIKO, MORIMOTO KENGO, TAKEDA HIDEAKI, SHIMOMURA YOSHIKI  人工知能学会全国大会論文集(CD−ROM)  20-  1  -4  2006  [Not refereed][Not invited]
  • Nakanishi Masatoshi, Suzuki Ryo, Shimomura Yoshiki, Takeda Hideaki, Yoshioka Masaharu  Proceedings of JSPE Semestrial Meeting  2006-  (0)  109  -110  2006  [Not refereed][Not invited]
     
    We argue that abduction for integrating theories can be a basic principle to formalize creative design processes. Based on this principle, we have proposed Universal Abduction Studio, a design environment in which designers combine different theories to arrive at better design. In this new approach to computational support of conceptual design, the system should offer various types of abductive reasoning from which designers can select an interesting design method. This paper aims at proposing the method of handling terminology by general common ontology and managing in the multiple. We also discuss technologies to implement proposed method of managing terminology.
  • Overview of the NTCIR-5 WEB Query Term Expansion Subtask. In Proceedings of the Fifth NTCIR Workshop Meeting on Evaluation of Information Access Technologies, National Institute of Informatics, 2005.
    Masaharu Yoshioka  Information Retrieval, Question Answering and Cross-Lingual Information Access  443  -454  2005/12  [Not refereed][Not invited]
  • YOSHIOKA Masaharu  IPSJ SIG Notes  81-  (108)  53  -60  2005/11/02  [Not refereed][Not invited]
     
    Query term expansion is a technique to modify initial query by adding some terms that represents users' information need for achieving higher retrieval performance. However, a good framework for evaluating this technique in isolation does not exist because the effect of query expansion depends on several issues. In this paper, I enumerate several features that may affect the quality of this technique and analyze retrieval experiment results by using these features.
  • Yoshioka Masaharu  Digital libraries  (29)  55  -62  2005/11  [Not refereed][Not invited]
  • YOSHIOKA Masaharu  IPSJ SIG Notes  136-  (42)  25  -30  2005/09/19  [Not refereed][Not invited]
     
    In many cases, it is not easy for a user of Information Retrieval (IR) to select appropriate query terms for representing his/her information need. Since this mismatch between query terms and information need affects the topic difficulty, it is better to have a framework to evaluate the effectiveness of the query terms. Therefore, in this paper, I propose some feature quantities that characterize quality of initial query in test collection. I also analyze correlation coefficient between these quantities and retrieval performance by using NTCIR-4 web test collection.
  • 山田哲弥, 梅田靖, 戸川契, 吉岡真治  IMS研究成果講演論文集  2005-  57-60  2005/07/20  [Not refereed][Not invited]
  • Shinichi Minato, Masaharu Yoshioka  情報処理学会研究報告. UBI, [ユビキタスコンピューティングシステム]  2005-  (60)  497  -501  2005/06/09  [Not refereed][Not invited]
     
    For tracing or finding the lost goods, we need active-type of RFID tags with built-in batteries, but those active tags have significant drawbacks such as the higher hardware cost and the finite battery lifetime. In this paper, we propose "double-decker" RFID tags, which can be separated into the two parts of ID tags. The new type of RFID tag is useful and cost-effective for tracing/keeping of a large number of goods. We present the basic structures of the double-decker tags and show some typical applications.
  • Yoshioka Masaharu  IPSJ SIG Notes  2005-  (42)  25  -30  2005/05/19  [Not refereed][Not invited]
     
    In many cases, it is not easy for a user of Information Retrieval (IR) to select appropriate query terms for representing his/her information need. Since this mismatch between query terms and information need affects the topic difficulty, it is better to have a framework to evaluate the effectiveness of the query terms. Therefore, in this paper, I propose some feature quantities that characterize quality of initial query in test collection. I also analyze correlation coefficient between these quantities and retrieval performance by using NTCIR-4 web test collection.
  • 吉岡 真治, Masaharu Yoshioka  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  20-  (3)  268  -268  2005/05/01  [Not refereed][Not invited]
  • 吉岡 真治  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  20-  (3)  268  -268  2005/05/01  [Not refereed][Not invited]
  • 吉岡 真治  人工知能  20-  (3)  268  -268  2005  [Not refereed][Not invited]
  • MORIMOTO Kengo, ONIKI Wataru, SHIMOMURA Yoshiki, YOSHIOKA Masaharu, TAKEDA Hideaki, UEDA Kanji  The Proceedings of Design & Systems Conference  2005-  (0)  168  -171  2005  [Not refereed][Not invited]
     
    It is important to unify different knowledge groups, which exist independently, and to extend knowledge gradually in order to obtain creative design solutions that are not obtained only with the existing design knowledge. With the aim of development of "Universal Abduction Studio", which is a system supporting creative design, a knowledge representation and a matching algorithm for creating hypothetic knowledge is proposed in this paper.
  • YOSHIOKA Masaharu  The Proceedings of Design & Systems Conference  2005-  (0)  334  -337  2005  [Not refereed][Not invited]
     
    Program management aims to support a management for a large and complex project. We have already proposed a framework for supporting this management process by reviewing a program from three different aspects (i.e., process planning, related information, and stake holder networks). However, this system only supports to represent and review a program by using manually inputted information. In this paper, I proposed a system that extract stakeholders information from a text to support constructing stake holder network. In this system, we categorize stakeholders in two types. One is collaborator and the other is opponent. For extracting these two types of stakeholders, we collected typical expression for describing them and made pattern match rule for extraction. For evaluating our methodology, we implemented the prototype system that uses Japanese newspaper articles as a resource for extraction.
  • ONIKI Wataru, MORIMOTO Kengo, SHIMOMURA Yoshiki, YOSHIOKA Masaharu, TAKEDA Hideaki  The proceedings of the JSME annual meeting  2005-  (0)  249  -250  2005  [Not refereed][Not invited]
     
    When a designer confronts a difficult design problem which cannot be solved only with his/her own knowledge, it is necessary to amplify it by introducing new perspectives and ideas for design. The authors have been developing a integrated computer aided design system for creative design. In this report, the authors propose a knowledge description framework which enables computers to analogize a bit of knowledge to another in a different field. Additionally, a tagging method of knowledge by RDF/XML and the implement of an editing tool for the method are mentioned in this report.
  • 吉田 守泰, 吉岡 真治  人工知能学会全国大会論文集  5-  (0)  95  -95  2005  [Not refereed][Not invited]
     
    大規模で複雑な事業などの支援を目的としたプログラムマネジメントにおいて、関係する利害関係者を理解することは重要な意味を持つ。本研究では、この情報を、新聞記事などのテキストからパタンを用いて抽出・分類する方法を提案する。
  • MORIMOTO KENGO, ONIKI WATARU, FUJIMOTO YUTAKA, SHIMOMURA YOSHIKI, YOSHIOKA MASAHARU, TAKEDA HIDEAKI  人工知能学会全国大会論文集(CD−ROM)  5-  (0)  282  -282  2005  [Not refereed][Not invited]
     
    本稿では，設計における問題解決に有効な仮説的知識を生成するために，概念ネッワークから知識モデルを生成し，そのモデルに対してアナロジーを適用する手法を提案する．さらに実験により本手法の有効性を検証する．
  • YOSHIDA MORIYASU, YOSHIOKA MASAHARU  人工知能学会全国大会論文集(CD−ROM)  19-  1  -4  2005  [Not refereed][Not invited]
  • Morimoto Kengo, Oniki Wataru, Fujimoto Yutaka, Shimomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2005-  (0)  183  -183  2005  [Not refereed][Not invited]
     
    It is important to unify different knowledge groups, which exist independently, and to extend knowledge gradually in order to obtain creative design solutions that are not obtained only with the existing design knowledge. Authors have proposed some knowledge representations for realizing knowledge extension gradual with the aim of development of “Universal Abduction Studio”, which is a system supporting creative design. In this paper, matching algorithm based on construction with deep case and validation checking algorithm with context are proposed.
  • Oniki Wataru, Morimoto Kengo, Fujimoto Yutaka, Shimomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2005-  (0)  182  -182  2005  [Not refereed][Not invited]
     
    When a designer confronts a difficult design problem which cannot be solved only with his/her own knowledge, it is necessary to amplify it by introducing new perspectives and ideas for design. The authors have been developing a integrated computer aided design system for creative design, and have proposed “knowledge scheme” to manage knowledge operations based on the meaning of knowledge in the last report. In this report, the authors propose a knowledge description format based on the knowledge scheme in the form of RDF/XML. In addition, the implementation of a knowledge editor to support the knowledge description is mentioned.
  • 森本憲悟, 下村芳樹, 藤本裕, 武田英明, 吉岡真治  精密工学会大会学術講演会講演論文集  2004-  A68  2004/09/01  [Not refereed][Not invited]
  • 梅田靖, 田辺繁彦, 戸川契, 吉岡真治  IMS研究成果講演論文集  2004-  101-104  2004/07/28  [Not refereed][Not invited]
  • 吉岡真治, 五十嵐優  人工知能学会人工知能基本問題研究会資料  56-  (0)  91  -96  2004/07/25  [Not refereed][Not invited]
  • HARAGUCHI Makoto, YOSHIOKA Masaharu  IPSJ SIG Notes  162-  (73)  7  -10  2004/07/15  [Not refereed][Not invited]
     
    We present here a method for constructing a common abstract event sequence from two or more documents also regarded as event sequences. To reduce the computational cost and to improve the precision of relevance judgement made by the abstract event sequences, we introduce a notion of base queries and consider a retrieval process to find an instance of base query that is similar to a given family of instance documents. By the base query and its instance documents, we can design a divide and conquer strategy for the construction of abstract common event sequence from the instances.
  • KONDO SHINSUKE, ISHIGAMI YOSHIAKI, UMEDA YASUSHI, SHIMOMURA YOSHIKI, YOSHIOKA MASAHARU  Designシンポジウム講演論文集  2004-  227-230  2004/07/08  [Not refereed][Not invited]
  • FUJIMOTO YUTAKA, SHIMOMURA YOSHIKI, TAKEDA HIDEAKI, YOSHIOKA MASAHARU  Designシンポジウム講演論文集  2004-  99-104  2004/07/08  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, HARAGUCHI Makoto  IPSJ SIG Notes  74-  (36)  87  -93  2004/03/26  [Not refereed][Not invited]
     
    It is not easy for a user of Information Retrieval (IR) system to select appropriate keywords. In this research, we proposed a new IR system that uses adaptive generalization of keywords. When the system can selects appropriate generalization by estimating user's intent, the system can generate good keywords that have high readability and good retrieval performance. In this report, we proposed new IR user interface that displaies generalized concepts by using relevant documents information. In addition, we propose a method to construct Boolean retrieval formula by using these generalized concepts and apply this method for Web information retrieval.
  • 吉岡 真治, 原口 誠  情報処理学会研究報告デジタルドキュメント（DD）  2004-  (36)  87  -93  2004/03/26  [Not refereed][Not invited]
     
    情報検索システムを利用する検索者にとって、適切なキーワードを選択することは必ずしも容易なことではない。本研究では、検索者にも理解しやすい概念階層の汎化という考え方を利用して、ユーザの検索意図を明示化すると共に、精度落ちを抑えた情報検索システムを提案している。本報では、汎化概念の関連文書における網羅性に注目し、検索に役に立つと考えられる汎化概念を明示化するインタフェースの作成について報告する。また、この網羅性に注目することにより、検索意図をより明確に表現するBoolean検索式の構築の支援とWeb検索への応用についても述べる。It is not easy for a user of Information Retrieval (IR) system to select appropriate keywords. In this research, we proposed a new IR system that uses adaptive generalization of keywords. When the system can selects appropriate generalization by estimating user's intent, the system can generate good keywords that have high readability and good retrieval performance. In this report, we proposed new IR user interface that displaies generalized concepts by using relevant documents information. In addition, we propose a method to construct Boolean retrieval formula by using these generalized concepts and apply this method for Web information retrieval.
  • 吉岡真治, 原口誠  言語処理学会年次大会発表論文集  10th-  (CD-ROM)  A10C2-01  2004/03/16  [Not refereed][Not invited]
  • 吉岡 真治  人工知能  19-  (1)  127  -128  2004  [Not refereed][Not invited]
  • YOSHIOKA Masaharu  Working Notes of the Fourth NTCIR Workshop Meeting, 2004  9  -16  2004  [Not refereed][Not invited]
  • MORIMOTO Kengo, FUJIMOTO Yutaka, SHIMOMURA Yoshiki, TAKEDA Hideaki, YOSHIOKA Masaharu  The Proceedings of Design & Systems Conference  2004-  (0)  332  -335  2004  [Not refereed][Not invited]
     
    In order to obtain a creative design solution, which is not obtained only with existing design knowledge, it is required to unify different knowledge groups, which exist independently, and to perform gradual extension of knowledge. Authors have proposed some knowledge representations and algorithms of knowledge integration for realizing knowledge extension with the aim of development of "Universal Abduction Studio", which supports creative design synthetically. In this paper, the fundamental framework for dealing with information as knowledge is proposed. And the representation and the knowledge integration based on it are proposed.
  • 武田 英明, 藤本 裕, 吉岡 真治, 下村 芳樹  人工知能学会全国大会論文集  4-  (0)  287  -287  2004  [Not refereed][Not invited]
     
    設計情報を計算機によって知的処理させるためには，テキストに機械処理可能なタグ付けをする必要がある．設計の場合，多様な文脈があるため，一意のタグづけは難しい．そこで本研究では，複数の解釈が可能なタグ付け方法を提案し，書籍の文章を例題にタグづけを試験的におこなった．
  • Morimoto Kengo, Shimomura Yoshiki, Fujimoto Yutaka, Takeda Hideaki, Yoshioka Masaharu  Proceedings of JSPE Semestrial Meeting  2004-  (0)  45  -45  2004  [Not refereed][Not invited]
     
    既存の設計知識だけでは得られないような創造的な設計解を得るためには、独立<br>に存在する異なる知識群を統合し、知識の段階的拡張を行うことが必要である。<br>著者らは創造的設計を総合的に支援する環境・UASの構築を目指し、これまでに<br>段階的な知識拡張を実現するための、複数の知識表現形式や知識統合の手法を提<br>案してきた。本報では自然言語などによって記述された情報を知識として取扱う<br>ための基本的枠組みを提案する。
  • 吉岡 真治, 五十嵐 優  人工知能学会全国大会論文集  4-  (0)  130  -130  2004  [Not refereed][Not invited]
     
    本研究では、ユーザが持つ問題に対し、類似事例などを提示しながら、問題解決を支援する知識マネージメントシステムの研究を行っている。本報では、オントロジーに基づく概念定義を利用した類似事例の検索と再利用の方法について述べる。
  • TAKEDA HIDEAKI, FUJIMOTO YUTAKA, YOSHIOKA MASAHARU, SHIMOMURA YOSHIKI  人工知能学会全国大会論文集(CD−ROM)  18-  1  -3  2004  [Not refereed][Not invited]
  • 石上佳照, 近藤伸亮, 梅田靖, 下村芳樹, 吉岡真治  Proceedings of JSPE Semestrial Meeting  2004-  (0)  297  -297  2004  [Not refereed][Not invited]
     
    アップグレード設計は、単に製品の物理的寿命を延長するだけでなく、機能的寿命を延長する設計手法である。本稿では、機能変更を伴う(e.g. 片面コピー機に二世代目で両面コピー機能を追加)アップグレード設計を扱う。従来、著者らが提案してきたアップグレード設計方法論に基づき、設計支援システムを開発し、実行例を述べ、本方法論の有効性を検証する。
  • Fujimoto Yutaka, Nomaguchi Yutaka, Simomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2004-  (0)  197  -197  2004  [Not refereed][Not invited]
     
    対象固有の領域知識を用いるだけでは、既存の設計解を改良することは可能であっても創造性の高い設計を支援することは困難である。即ち、創造性の高い設計を支援するためには、異なる領域に属する知識を互いに関連づけることによる問題の再定義や、新しい知識を作成するための手段が必要である。本報では、前報ですでに提案したオントロジーを用いた知識表現の枠組みとその評価結果に基づき、より柔軟な知識利用を実現するための新しい知識表現の手法を提案するとともに、その有効性について検証する。
  • 吉岡 真治  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  19-  (1)  127  -128  2004/01/01  [Not refereed][Not invited]
  • 吉岡 真治  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  18-  (3)  367  -368  2003/05/01  [Not refereed][Not invited]
  • YOSHIOKA Masaharu  Journal of Japanese Society for Artificial Intelligence  18-  (2)  107  -107  2003/03/01  [Not refereed][Not invited]
  • Masaharu Yoshioka  Journal of the Japanese Society for Artificial Intelligence  18-  (2)  107  -107  2003
  • 吉岡 真治  人工知能  18-  (3)  367  -368  2003  [Not refereed][Not invited]
  • 吉岡 真治  人工知能  18-  (1)  101  -101  2003  [Not refereed][Not invited]
  • ISHIGAMI Yoshiaki, KONDOH Shinsuke, UMEDA Yasushi, SHIMOMURA Yoshiki, YOSHIOKA Masaharu  The proceedings of the JSME annual meeting  2003-  (0)  323  -324  2003  [Not refereed][Not invited]
     
    The objective of upgrade design is to extend product life so as to reduce consumption volume of resources and energy usage. However, difficulty of upgrade design resides within prediction of future trend such as technological trends and market trends, and upgrade design should be robusts against uncertainty of future predection. This paper proposes a design methodology for upgradable products, and illustrates a prototype upgrade design support tool which can handle changes of user requirement and market trend. With a design example of a vacuume cleaner, this papar also demonstrates the features of the upgrade design tool.
  • YOSHIOKA Masaharu, NOMAGUCHI Yutaka, FUJIMOTO Yutaka, SHIMOMURA Yoshiki, TAKEDA Hideaki  The Proceedings of Design & Systems Conference  2003-  (0)  189  -191  2003  [Not refereed][Not invited]
     
    Since previous designer may already generate possible combination of existing domain knowledge, it is difficult to design creative solution without creating innovative knowledge. We have already confirmed that combination of existing knowledge from different domain may useful for this process by analyzing design cases. In this paper, we propose methodology to generate hypothetical knowledge for creative design by combining existing knowledge.
  • 武田 英明, 坂井 宏充, 野間口 大, 吉岡 真治, 下村 芳樹, 冨山 哲男  人工知能学会全国大会論文集  3-  (0)  183  -183  2003  [Not refereed][Not invited]
     
    本研究の目的は、設計の創造過程を表現し、計算機上で支援する環境を構築することである。設計の創造的活動はアブダクション、特に理論を統合するアブダクションとしてモデル化することができる。理論を統合するアブダクションの実現には多様なアブダクションのスキーマが併せて利用されるが、ここではアナロジーに基づくアブダクションに注目して、アナロジーに基づく知識統合による設計支援システムのプロトタイプを構築した。
  • 原口 誠, 吉岡 真治, 溝江 彰人, 四ッ谷 雅輝  人工知能学会全国大会論文集  3-  (0)  189  -189  2003  [Not refereed][Not invited]
     
    類似文書群の中で初めて言葉の重要度と関連性が定まるとの立場から、物語性を考慮した文の重要度に基づく文書汎化手法を提案する。また、汎化の品質向上に寄与する辞書の更新についても実験的検当結果を報告する。
  • 吉岡 真治, 五十嵐 優, 社本 康広  人工知能学会全国大会論文集  3-  (0)  93  -93  2003  [Not refereed][Not invited]
     
    本研究では、ユーザが持つ問題に対し、類似事例などを提示しながら、問題解決を支援する知識マネージメントシステムの研究を行う。本報では、問題解決のタスクの種類に関する分析と、その分析に応じて適切な知識の構造化手法について提案する。
  • Yagi Hiroaki, Kondoh Shinsuke, Umeda Yasushi, Shimomura Yoshiki, Yoshioka Masaharu  Proceedings of JSPE Semestrial Meeting  2003-  (0)  257  -257  2003  [Not refereed][Not invited]
     
    製品のアップグレード設計とは、モジュールを追加、交換することにより製品の機能をアップグレードし機能寿命の延命化を図るものであり、機能の陳腐化による廃棄行動を抑制可能である。本稿では、アップグレード時に機能の追加／削除を可能にするアップグレード製品設計の方法論を提案する。具体的には、対象製品の機能&middot;挙動&middot;構造の関係を操作することにより、機能の追加／削除が可能である、追加モジュールおよび世代間で共通のプラットフォームを決定する。
  • 松田 暁, 下村 芳樹, 吉岡 真治, 梅田 靖  Proceedings of JSPE Semestrial Meeting  2003-  (0)  258  -258  2003  [Not refereed][Not invited]
     
    著者らは，製品の機能的な寿命の延長を実現するための方法であるアップグレード設計方法論の研究を行っている。本研究は、予測された市場の技術動向に対し追従可能であるよう，製品に施す機能的なアップグレードを計画し、その計画を最適に実現可能であるように製品を設計するための方法を与えることを目的とするものであり、製品のアップグレードを効果的に行うためには、製品アップグレードの計画段階でそれによる効果を把握できることが好ましい。本報では、市場に投入される技術の内容、タイミング、製品使用者の指向等に影響される不確実性を考慮した上で、消費者要求の将来予測を行い、それに基づいた製品ラインナップに要求される仕様を把握する方法について提案する。
  • 下村 芳樹, 坂井 宏充, 野間口 大, 吉岡 真治, 武田 英明, 冨山 哲男  Proceedings of JSPE Semestrial Meeting  2003-  (0)  27  -27  2003  [Not refereed][Not invited]
     
    著者らは設計における創造性の理解とその計算機による支援を目的とする研究を行っており、その一環として統合的推論環境Universal Abduction Studioの開発に着手している。創造性の高い設計支援を実現する一つの鍵は、対象固有の領域知識のみでなく、異なる領域に属する知識を横断的に利用可能とすることであると考えられる。本研究では、異領域の知識を横断的に利用する推論形態としてのAbductionに着目し、これを統合的に実装することにより創造性の高い設計支援のための推論環境を提供するUniversal Abduction Studioの提案を行う。
  • Sakai Hiromitsu, Nomaguchi Yutaka, Shimomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2003-  (0)  28  -28  2003  [Not refereed][Not invited]
     
    著者らは設計における創造性の理解とその計算機による支援を目的に研究を行っている。創造性の高い設計の支援を考える場合、対象固有の領域知識に留まらず、異なる領域に属する知識を横断的に利用可能とすることが必要である。本研究では、知識間の構造的な類似性を発見する枠組みと、これを創造性の高い設計支援に適用する手法を提案する。特に、知識間の類似性発見の手法について、例題よりその有効性を検証する。
  • 吉岡 真治, 五十嵐 優, 社本 康広  人工知能学会全国大会論文集  17-  1  -3  2003  [Not refereed][Not invited]
  • 原口 誠, 吉岡 真治, 溝江 彰人  人工知能学会全国大会論文集  17-  1  -3  2003  [Not refereed][Not invited]
  • Fujimoto Yutaka, Nomaguchi Yutaka, Shimomura Yoshiki, Yoshioka Masaharu, Takeda Hideaki  Proceedings of JSPE Semestrial Meeting  2003-  (0)  39  -39  2003  [Not refereed][Not invited]
     
    対象固有の領域知識を用いるだけでは、既存の設計解を改良することは可能であっても創造性の高い設計を支援することは困難である。即ち、創造性の高い設計を支援するためには、異なる領域に属する知識を関連づけることによる問題の再定義や、新しい知識を作成するための手段が必要である。本報では、前報で提案した知識間の類似性を発見する枠組とその評価結果に基づき、より高度な知識利用を実現するための知識表現の枠組みに対する要求仕様を明らかにするとともに、その具体的な実現方法に関して考察する。
  • 吉岡 真治  Journal of Japanese Society for Artificial Intelligence  18-  (1)  101  -101  2003/01/01  [Not refereed][Not invited]
  • 吉岡 真治  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  17-  (5)  649  -650  2002/09/01  [Not refereed][Not invited]
  • HARAGUCHI Makoto, NAKANO Shigetora, YOSHIOKA Masaharu  IPSJ SIG Notes  150-  (66)  99  -103  2002/07/15  [Not refereed][Not invited]
     
    Given two or more similar documents in the form of texts, we present a notion of maximal analogies representing maximal sequences consisting of pairs of similar events in the documents. They are required to satisfy certain cost condition so that meaningless similarities between documents are never concluded. A bottom-up search procedure to find a maximal analogy satisfying the cost condition is also presented. In addition to the set of similar documents, we suppose another set of dissimilar ones. Then a maximal analogy is furthermore tested for their appropriateness so as not to explain the latter ones. The test can be performed by an effective subsumption check procedure.
  • HARAGUCHI Makoto, NAKANO Shigetora, YOSHIOKA Masaharu  IEICE technical report. Natural language understanding and models of communication  102-  (200)  1  -5  2002/07/09  [Not refereed][Not invited]
     
    Given two or more similar documents in the form of texts, we present a notion of maximal analogies representing maximal sequences consisting of pairs of similar events in the documents. They are required to satisfy certain cost condition so that meaningless similarities between documents are never concluded. A bottom-up search procedure to find a maximal analogy satisfying the cost condition is also presented. In addition to the set of similar documents, we suppose another set of dissimilar ones. Then a maximal analogy is furthermore tested for their appropriateness so as not to explain the latter ones. The test can be performed by an effective subsumption check procedure.
  • YOSHIOKA Masaharu, HARAGUCHI Makoto, OHKUBO Yoshiaki  IPSJ SIG Notes  127-  (41)  151  -158  2002/05/21  [Not refereed][Not invited]
     
    It is not easy for a user of Information Retrieval (IR) system to select appropriate keywords. Therefore, many IR systems have capability to modify keywords by estimating user's intent. However, since modified keywords are usually represented as complicated form, it is difficult to judge the appropriateness of them. In this research, we proposed a new IR system that uses adaptive generalization of keywords. when the system can selects appropriate generalization by estimating user's intent, the system can generate good keywords that have high readability and good retrieval performance. In this report, we proposed to use the contribution value of each keyword for relevance judgement to select appropriate generalization. In addition, we confirm general concept structure stored in a thesaurus is not sufficient for representing particular user's intent. Therefore, we proposed a purpose oriented method to modify concept structure.
  • 吉岡 真治, 原口 誠, 大久保 好章  情報処理学会研究報告情報学基礎（FI）  2002-  (41)  151  -158  2002/05/21  [Not refereed][Not invited]
     
    情報検索システムを利用する検索者にとって、適切なキーワードを選択することは必ずしも容易なことではない。この問題に対し、検索者の検索意図を推定し、検索者の補完を行う検索システムが提案されている。しかし、補完した結果は複雑であることが多く、検索者が補完結果自体を評価することが困難である。本研究では、検索者にも理解しやすい概念階層の汎化という考え方を利用して、ユーザの検索意図を明示化すると共に、精度落ちを抑えた情報検索システムを提案する。本報では、概念階層の汎化のレベルを決定するために、検索者により入力された検索語が持つ適合性判定への寄与度を考える。また、本システムの性能を向上させるためには、概念階層自体が目的に応じて再構成されることが必要であることについて述べ、目的指向の概念階層の修正の方法を提案する。It is not easy for a user of Information Retrieval (IR) system to select appropriate keywords. Therefore, many IR systems have capability to modify keywords by estimating user's intent. However, since modified keywords are usually represented as complicated form, it is difficult to judge the appropriateness of them. In this research, we proposed a new IR system that uses adaptive generalization of keywords. When the system can selects appropriate generalization by estimating user's intent, the system can generate good keywords that have high readability and good retrieval performance. In this report, we proposed to use the contribution value of each keyword for relevance judgement to select appropriate generalization. In addition, we confirm general concept structure stored in a thesaurus is not sufficient for representing particular user's intent. Therefore, we proposed a purpose oriented method to modify concept structure.
  • TAKEUGHI KOICHI, UCHIYAMA KIYOKO, YOSHIOKA MASAHARU, KAGEURA KYO, KOYAMA TERUO  IPSJ Journal  43-  (5)  1446  -1456  2002/05/15  [Not refereed][Not invited]
     
    In this paper, we describe a principled approach for analyzing relations between constituent words of compound nouns, specifically those whose heads are deverbal nouns, based on the classification of deverbal nouns by their lexical conceptual structure (LCS) and the classification of nouns in modifier position vis-a-vis LGS of head deverbal nouns. There are two kinds of relations of compounds with head deverbal nouns. The one is that a modifier noun becomes an argument of deverbal head and the other is that a modifier becomes an adjunct. It is an important starting point for analyzing relat...
  • Masaharu Yoshioka, Noriko Kando  Journal of Japanese Society for Artificial Intelligence  17-  (3)  295  -295  2002/05/01  [Not refereed][Not invited]
  • 吉岡 真治  人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence  17-  (3)  372  -373  2002/05/01  [Not refereed][Not invited]
  • KURIYAMA KAZUKO, YOSHIOKA MASAHARU, KANDO NORIKO  情報処理学会論文誌. データベース  43-  (2)  48  -59  2002/03/15  [Not refereed][Not invited]
     
    The purposes of this study are to examine whether there is an effect on the relative evaluation of the IR systems using the relevance judgments of the test collection NTCIR-2 made by the pooling method and additional interactive searches, and to investigate whether the NTCIR-2 is effective for evaluating the IR systems, the search results of which were not used for the pooling. We carried out experiments using different lists of relevance judgments and search results submitted for the test of the 2nd NTCIR Workshop. First, we evaluated the search results using the list of the final relevanc...
  • 梅森 嘉匡, 八木 宏明, 近藤 伸亮, 梅田 靖, 下村 芳樹, 吉岡 真治  精密工学会大会学術講演会講演論文集  2002-  (1)  138  -138  2002/03/01  [Not refereed][Not invited]
  • 吉岡真治  2002年度人工知能学会全国大会講演論文集  2002
  • 吉岡 真治  人工知能  17-  (5)  649  -650  2002  [Not refereed][Not invited]
  • 吉岡 真治  Journal of the Japanese Society for Artificial Intelligence  17-  (3)  372  -373  2002  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, KANDO Noriko  Journal of the Japanese Society for Artificial Intelligence  17-  (3)  295  -295  2002  [Not refereed][Not invited]
  • OGAWA Yasushi, SASAKI Yutaka, MASUYAMA Shigeru, MURATA Masaki, YOSHIOKA Masaharu  Journal of the Japanese Society for Artificial Intelligence  17-  (3)  306  -311  2002  [Not refereed][Not invited]
  • ISHIGAMI Yoshiaki, KONDOH Shinsuke, UMEDA Yasushi, SHIMOMURA Yoshiki, YOSHIOKA Masaharu  The Proceedings of Design & Systems Conference  2002-  (0)  56  -59  2002  [Not refereed][Not invited]
     
    Upgrade design is one of the usefull approaches to realize inverse manufacturing because upgrade design can extend product lives and reduce consumption volume of resources and energy usage. However, difficulty of upgrade design resides within prediction of future trend such as technological trends and market trends. This papar propose a design method for upgradable products, and prototype design support tool which can be adapted changes of user requirement and market trend. Using design example of vacuume cleaner, this papar demonstrates that this support tools is effective for designing upgradable products.
  • SAKAI Hiromitsu, NOMAGUCHI Yutaka, YOSHIOKA Masaharu, TAKEDA Hideaki, SHIMOMURA Yoshiki, TOMIYAMA Tetsuo  The Proceedings of Design & Systems Conference  2002-  (0)  410  -411  2002  [Not refereed][Not invited]
     
    Although design support systems have been well developed in geometric and detail design areas, they are still not successful in conceptual or creative design domain. In our opinion, the main difficulty is knowledge and its operation that play a crucial role in conceptual design. In this paper, we propose "Universal Abduction Studio" that is a new approach to support designers by computers, and discuss technologies to realize it.
  • Yoshioka Masaharu, Igarashi You, Shamoto Yasuhiro  Proceedings of the Annual Conference of JSAI  2-  (0)  142  -142  2002  [Not refereed][Not invited]
  • 吉岡 真治, 五十嵐 優, 社本 康広  人工知能学会全国大会論文集  16-  1  -4  2002  [Not refereed][Not invited]
  • Nomaguchi Yutaka, Yoshioka Masaharu, Tomiyama Tetsuo  Transactions of the Japanese Society for Artificial Intelligence  17-  (17)  104  -113  2002  [Not refereed][Not invited]
     
    In this paper, we propose a fundamental idea of a new CAD mechanism to facilitate design knowledge management. This mechanism encourages a designer to externalise his/her knowledge during a design process and facilitates sharing and reuse of such externalised design knowledge in later stages. We also describe the implementation of this idea called DDMS (Design Documentation Management System). DDMS works as a front end to KIEF (Knowledge Intensive Engineering Framework), which we have been developing. We also illustrate an example of machining tool design to demonstrate the features of DDMS.
  • KURIYAMA Kazuko, YOSHIOKA Masaharu, KANDO Noriko  IPSJ SIG Notes  63-  (74)  9  -16  2001/07/24  [Not refereed][Not invited]
     
    The purpose of this study is to examine whether there is an effect on the relative evaluation of the IR systems using the relevance judgments made by the pooling method and additional interactive searches. We carried out experiments using different lists of relevance judgments and search results submitted for the test of the 2nd NTCIR Workshop. First, we evaluated the search results using the list of the final relevance judgments F of NTCIR-2 and F-I, that is, the F without the unique relevant documents found by the additional interactive searches I. Second, we made pools from the search results for each of the sub-tasks and evaluated the search results using the pools as lists of relevance judgments. Almost the same rankings of the search results were produced by using the pools as lists of relevance judgments for system evaluation. Therefore our results verified the reliability of test collection as an evaluation tool, which was based on pooling method.
  • KANDO Noriko, KURIYAMA Kazuko, YOSHIOKA Masaharu  IPSJ SIG Notes  63-  (74)  105  -112  2001/07/24  [Not refereed][Not invited]
     
    This paper proposes new measures, weighted R precision and wighted average precision, for information retrieval(IR) system testing with ranked output based on multi-grade relevance judgments. Relevance judgments are one of the most fundamental factors in evaluation of search effectiveness of IR systems. Intuitively, the highly relevant documents are more important for users than partial relevant ones and the documents retrieved in the higher ranks in the ranked list are more important. Therefore the systems producing the search results in which higher relevant documents in higher ranks in the ranked list should be rated as better thought relevance judgments must be done in binary to calculate traditional IR system evaluation measures "recall" and "precision". Based on the review of existing IR system evaluation measures, we decided that either of proposed measures is single number and averageable over number of runs. The paper then shows the evaluation results of the Japanese IR task at the second NTCIR Workshop, in which four-grade relevance judgments, highly relevant, relevant, partially relevant, and irrelevant, were used. Finally implications of "partial relevance" are discussed in the relation to user-oriented IR system evaluation studies.
  • The effect of cross-lingual pooling on evaluation
    Kuriyama,K, Yoshioka, M, Kando, N  Proceedings of the Second NTCIR Workshop on Research in Chinese & Japanese Text Retrieval and Text Summarization., Tokyo, Japan, May 2000-March 2001, National Institute of Informatics  297  -310  2001/07  [Not refereed][Not invited]
  • Kando, N, Kuriyama, K, Yoshioka, M  Proceedings of the Second NTCIR Workshop on Research in Chinese & Japanese Text Retrieval and Text Summarization, Tokyo, Japan, May 2000-March 2001, National Institute of Informatics  73  -96  2001/07  [Not refereed][Not invited]
  • Yoshioka, M, Kuriyama, K, Kando, N  Proceedings of the Second NTCIR Workshop on Research in Chinese & Japanese Text Retrieval and Text Summarization., Tokyo, Japan, May 2000-March 2001, National Institute of Informatics  291  -296  2001/07  [Not refereed][Not invited]
  • 下村芳樹, 吉岡真治, さい剛民, 梅森嘉匡, 梅田靖  精密工学会大会学術講演会講演論文集  2001-  365  2001/03/05  [Not refereed][Not invited]
  • 人と情報システムとのシンビオシスのための情報検索および情報活用支援に関する研究 -- Web文書検索を中心として--
    栗山和子, 吉岡真治, 江口浩二, 岸田和明, 神門典子, 影浦峡, 高野明彦, 大山敬三, 安達淳  COEプロジェクト平成12年度報告書『シンビオティック情報システムへの概念形成と実現技術』, 国立情報学研究所,2001  47  -54  2001/03  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, Takeda Hideaki, WASHIO Takashi, TOMIYAMA Tetsuo  The Proceedings of Design & Systems Conference  2001-  (0)  281  -284  2001  [Not refereed][Not invited]
     
    Since design knowledge plays a crucial role in design process, we are conducting research on design reasoning model that takes notice of design knowledge. To make a design reasoning model, we have already proposed a reasoning framework of design and analyzed basic operations for the framework. However, since these operations are just fragments of whole design process, it is necessary to describe design operational knowledge that can control the entire reasoning operations by constructing the sequence of these fragments. So, in this paper, we propose a synthesis language for describing design operational knowledge.
  • Nomaguchi Yutaka, Tsumaya Akira, Yoshioka Masaharu, Washio Takashi, Takeda Hideaki, Murakami Tamotsu, Tomiyama Tetsuo  The Proceedings of Design & Systems Conference  2001-  (0)  285  -288  2001  [Not refereed][Not invited]
     
    Since design knowledge plays a crucial role in design process, we are conducting research on design reasoning model that takes notice of design knowledge. To make a design reasoning model, we have already proposed a reasoning framework of design and analyzed basic operations for the framework, and then propose a synthesis language for describing design operational knowledge. In this paper, we propose our implementation of this model, and verify it to compare design processes replayed in the system with ones of actual design sessions.
  • TSUMAYA Akira, NOMAGUCHI Yutaka, YOSHIOKA Masaharu, TAKEDA Hideaki, MURAKAMI Tamotsu, TOMIYAMA Tetsuo  The Proceedings of Design & Systems Conference  2001-  (0)  289  -292  2001  [Not refereed][Not invited]
     
    In engineering design research, numerous design process models have already been proposed. For constructing model of synthesis, it is very important to know the position of the model among them. This paper reports results of comparison of some design process models with the synthesis oriented thought process model. We use a reference model as an example of design, compare it with some existing design process models, and the results of this analysis are described.
  • NOMAGUCHI Yutaka, YOSHIOKA Masaharu, TOMIYAMA Tetsuo  設計シンポジウム講演論文集  2001-  (19)  109  -116  2001  [Not refereed][Not invited]
     
    In this paper, we propose a fundamental idea of a new CAD architecture to facilitate design knowledge management. This architecture is based on knowledge-based design environment for encouraging the designer to externalize his/her knowledge. We also discuss the implementation of this idea. DDMS (Design Documentation Management System) works as a front end for KIEF (Knowledge Intensive Engineering Framework), which we have been developing. We also illustrate an example of machining tool design to demonstrate the system's features.
  • Sakai Hiromitsu, Nomaguchi Yutaka, Yoshioka Masaharu, Shimomura Yoshiki, Tomiyama Tetsuo  The Proceedings of Design & Systems Conference  2001-  (0)  204  -206  2001  [Not refereed][Not invited]
     
    In engineering activities, management of designers' knowledge is becoming crucial to share and reuse it efficiently. Toward this issue, we are studying a methodology to capture the designer's thought process, and to semi-automatically arrange it into a refined document, which addresses a particular objective with particular thought patterns. The key issues of this methodology are to acquire the thought patterns and to define relations among them. This paper reports our algorithm to acquire thought patterns from a record of designer's thought processes and any other types of documents including patent documents and academic papers. We employ a knowledge discovery approach which is recently discussed in data mining research.
  • UMEMORI Yoshimasa, KONDOH Shinsuke, UMEDA Yasushi, SHIMOMURA Yoshiki, YOSHIOKA Masaharu  The Proceedings of Design & Systems Conference  2001-  (0)  269  -272  2001  [Not refereed][Not invited]
     
    Conventional products tend to be thrown away because of functional obsoleteness before reaching the end of physical life of the products. This is one of the main causes of the mass disposal problem. In order to solve this problem, this paper proposes a design methodology for upgradable products. Especially, this paper discusses a method dealing with uncertainty of long-term planning in the design methodology. The proposed method deals with uncertainty as ranges of parameter values, and derives a design solution that realizes required upgrades, while adapting to estimated uncertainty. This method makes a design object robust and tolerant against the uncertainty and, therefore, makes design for upgradability more feasible.
  • YOSHIOKA Masaharu, SHAMOTO Yasuhiro  The Proceedings of Design & Systems Conference  2001-  (0)  321  -324  2001  [Not refereed][Not invited]
     
    In engineering activities, engineers solve various kinds of problems. Most of these problems are routine problems that are easy to formalize and formalized knowledge is useful for these problems. On the other hand, there are complicated problems that is difficult to formalize all of the problem solving process. One of the useful information to solve such complicated problems is result of the similar problem solving process. In addition, later in such complicated problem solving process, most of the decomposed subproblems are able to handle as routine problems. In this research, we decide to use formalized knowledge and document information that represents the history of problem solving processes to support to solve both types of the problem. For that purpose, we propose a new knowledge management system that supports a user to solve the problem through document processing. In this support process, the system retrieves related documents for user's problems by using formalized knowledge and information retrieval techniques. After retrieving related documents, he/she modifies them to fit his/her problems. Modified information is associated with the formalized knowledge and stored as documents for future reuse.
  • Takeda Hideaki, Yoshioka Masaharu, Tomiyama Tetsuo  Proceedings of the Annual Conference of JSAI  1-  (0)  30  -30  2001  [Not refereed][Not invited]
  • 武田 英明, 吉岡 真治, 冨山 哲男  人工知能学会全国大会論文集  15-  1  -4  2001  [Not refereed][Not invited]
  • KURIYAMA Kazuko, YOSHIOKA Masaharu, KANDO Noriko  IPSJ SIG Notes  60-  (91)  9  -16  2000/09/27  [Not refereed][Not invited]
     
    In this paper we describe large scale test collections, NTCIR-1 and NTCIR-2, and the 2nd NTCIR Workshop. We also show differences in the records of the test collections from point of view of statistical data ; length of the records, their distribution, number of different terms, their frequencies, and so on.
  • Kuriyama Kazuko, Yoshioka Masaharu, Kando Noriko  Digital libraries  18-  (18)  57  -64  2000/09  [Not refereed][Not invited]
  • 栗山和子, 吉岡真治, 神門典子, Kuriyama kazuko, Yoshioka Masaharu, Kando Noriko, クリヤマカズコ, ヨシオカマサハル, カンドノリコ  ディジタル図書館  (18)  57  -64  2000/09  [Not refereed][Not invited]
  • 吉岡 真治, 武田 英明, 妻屋 彰, 村上 存, 冨山 哲男  精密工学会大会学術講演会講演論文集  2000-  (2)  259  -259  2000/09/01  [Not refereed][Not invited]
  • 野間口 大, 妻屋 彰, 吉岡 真治, 武田 英明, 村上 存, 冨山 哲男  精密工学会大会学術講演会講演論文集  2000-  (2)  260  -260  2000/09/01  [Not refereed][Not invited]
  • TAKEDA Hideaki, YOSHIOKA Masaharu, TOMIYAMA Tetsuo  Proceedings of the Annual Conference of JSAI  14-  (0)  464  -467  2000/07/03  [Not refereed][Not invited]
  • YOSHIOKA Masaharu, TAKEDA Hideaki, TOMIYAMA Tetsuo  Proceedings of the Annual Conference of JSAI  14-  (0)  468  -471  2000/07/03  [Not refereed][Not invited]
  • TAKEUCHI Koichi, UCHIYAMA Kiyoko, YOSHIOKA Masaharu, KAGEURA Kyo, KOYAMA Teruo  IPSJ SIG Notes  57-  (29)  71  -78  2000/03/21  [Not refereed][Not invited]
     
    The aim of our research is to build a Japanese compound analyzer using grammatical constraint. To analyze the compounds which have verbs in their heads is to reveal the relationship between incorporated nouns and the head verbal nouns. The kind of relationship is adjunctive and internal argument of a head verb. In our previous research, we categorize verbs based on their argument structure and lexical conceptual structure (LCS), but it is necessary to categorize nouns to reveal the relationship between a noun and a head verb in compounds. Therefore we propose the method to categorize nouns based on argument structure and LCS of head verbs in this paper. In the case of categorization, we also use qualia structure of generative lexicon. After the assumptions of restriction about categorized nouns is constructed, we attempt the experiment of analyzing technical terms (103 words) in the domain of information processing. Finally we show the experimental results that 92 words (89%) can be solved using our restriction and discuss the limitation of our model.
  • Takeuchi Koichi, Uchiyama Kiyoko, Yoshioka Masaharu, Kageura Kyo, Koyama Teruo  IPSJ SIG Notes  2000-  (29)  71  -78  2000/03/21  [Not refereed][Not invited]
     
    The aim of our research is to build a Japanese compound analyzer using grammatical constraint. To analyze the compounds which have verbs in their heads is to reveal the relationship between incorporated nouns and the head verbal nouns. The kind of relationship is adjunctive and internal argument of a head verb. In our previous research, we categorize verbs based on their argument structure and lexical conceptual structure (LCS), but it is necessary to categorize nouns to reveal the relationship between a noun and a head verb in compounds. Therefore we propose the method to categorize nouns based on argument structure and LCS of head verbs in this paper. In the case of categorization, we also use qualia structure of generative lexicon. After the assumptions of restriction about categorized nouns is constructed, we attempt the experiment of analyzing technical terms (103 words) in the domain of information processing. Finally we show the experimental results that 92 words (89%) can be solved using our restriction and discuss the limitation of our model.
  • Takeuchi Koichi, Uchiyama Kiyoko, Yoshioka Masaharu, Kageura Kyo, Koyama Teruo  Research bulletin of the National Center for Science Information Systems  12-  (12)  7  -15  2000/03  [Not refereed][Not invited]
     
    In the previous research of analyzing compound words, the technique using statistics of word co-occurrence is well known. However, since some grammatical and semantic restrictions must control the construction of compound words, only the statistical technique is not enough to analyze compound words. In this paper, we analyze technical terms that consist of two nouns whose head is verbal noun by means of the rich lexical knowledge that are a level ordering assumption, an argument structure and a lexical conceptual structure used on morphology or syntax. We will show the following three results: 1) there are only two kinds of relationship that is a modifier or a argument for a verbal noun in the relationship between two words in a compound word, 2) the analysis of compound words depends on argument structure of verbal noun, and 3) the structure of some compound words are changed according to the context. Finally we discuss constructing compound words analyzer from the results.
  • 吉岡 真治, 妻屋 彰, 冨山 哲男, 武田 英明  精密工学会大会学術講演会講演論文集  2000-  (1)  18  -18  2000/03/01  [Not refereed][Not invited]
  • 武田 英明, 吉岡 真治, 妻屋 彰, 冨山 哲男  精密工学会大会学術講演会講演論文集  2000-  (1)  19  -19  2000/03/01  [Not refereed][Not invited]
  • 妻屋 彰, 野間口 大, 吉岡 真治, 武田 英明, 村上 存, 冨山 哲男  精密工学会大会学術講演会講演論文集  2000-  (1)  20  -20  2000/03/01  [Not refereed][Not invited]
  • 今川 晴紀, 下村 芳樹, 吉岡 真治, 梅田 靖  精密工学会大会学術講演会講演論文集  2000-  (1)  96  -96  2000/03/01  [Not refereed][Not invited]
  • 下村芳樹, 今川晴紀, 吉岡真治, 梅田靖  日本機械学会設計工学・システム部門講演会講演論文集  9th-  501  -504  1999/11/19  [Not refereed][Not invited]
  • TAKEUCHI Koichi, UCHIYAMA Kiyoko, YOSHIOKA Masaharu, KAGEURA Kyo, KOYAMA Teruo  IEICE technical report. Natural language understanding and models of communication  99-  (387)  7  -14  1999/10/25  [Not refereed][Not invited]
     
    It is difficult problem to know how to prepare lexical knowledge in order to analyze Japanese compound words. In this paper, we analyze technical terms that consist of two verbal nouns by means of the rich lexical knowledge that are a level ordering assumption, an argument structure and a lexical conceptual structure used on morphology or syntax. Our analysis of compound words aims to reveal the relations between the constituent words of compound words. From the results, our method can explain some parts of phenomena of the impossible compound words and the conversion of verbal noun into noun in compound words. The lexical knowledge we have assumed is aimed at describing the details of verbal nouns. While analyzing verbal nouns, we also clarify how to describe the lexical knowledge of noun words.
  • Yoshioka Masaharu, Nomaguchi Yutaka, Tomiyama Tetsuo  IPSJ SIG Notes  99-  (88)  63  -68  1999/10/21  [Not refereed][Not invited]
     
    In a design process, the designer uses a variety of information and knowledge for generating design solution and evaluating it. These types of process information are useful not only for understanding the design rationale but also for the source of designers' knowledge. Therefore, many companies try to store these types of information as design documents. However, since most of the documents are made for explaining the design solution after finishing whole design, most of the documents do not cover all of information in design process. In addition, there is not good framework to retrieve and reuse these documents and that prevents the designers to reuse these useful documents. For tackling these problems, we have proposed the design documentation management system. This system can record design documents as a result of design process. The system can also retrieve and reuse related design documents based on the design context. In this report, we discuss the Japanization of the design documentation system and application of text retrieval technique.
  • UCHIYAMA Kiyoko, TAKEUCHI Koichi, YOSHIOKA Masaharu, KAGEURA Kyo, KOYAMA Teruo  Technical report of IEICE. Thought and language  99-  (237)  7  -12  1999/07/26  [Not refereed][Not invited]
     
    Technical compound nouns are often used as keywords for information retrieval. The mechanism how nominal compounds are generated can be clarified by analyzing internal structures of nominal compounds. The conventional approach has centered on the analysis of semantic relations among constituent elements of nominal compounds. We think structural and grammatical characteristics play the key role in nominal compounds. In our study, we propose original categories in terms of structural and grammatical characteristics, and examine modifications and connections among constituent elements of compound words, using the original categories.
  • Yoshioka Masaharu  Research bulletin of the National Center for Science Information Systems  11-  (11)  27  -31  1999/03  [Not refereed][Not invited]
     
    There are several types of research articles; e.g., survey articles, those that propose new technologies, those that apply existing technologies to unsolved problems. In this paper, I propose a methodology to categorize these types of research articles by analyzing roles of the author added keywords. First, I categorize verbs that are frequently used in research articles into five role categories. Second, I propose methodology to analyze roles of the keywords by using the categories of verbs. To validate this methodology, I apply it to the real research articles and explain its result.
  • Uchiyama Kiyoko, Takeuchi Koichi, Yoshioka Masaharu, Kageura Kyo, Koyama Teruo  Research bulletin of the National Center for Science Information Systems  11-  (11)  49  -57  1999/03  [Not refereed][Not invited]
     
    Grammatical information for constituent elements of nominal compounds are needed for analyzing nominal compounds, especially those of technical fields. In this research we analyzed the grammatical categories suitable for nominal compound analysis, as it is difficult to apply conventional grammatical categories defined with respect to words in sentences to the analysis of nominal compounds.
  • YOSHIOKA Masaharu, OKADA Maho, KAGEURA Kyo, KOYAMA Teruo  IPSJ SIG Notes  53-  (20)  41  -48  1999/03/01  [Not refereed][Not invited]
     
    We are currently constructing the terminology-motivated corpus (the NACSIS corpus) as a part of the NACSIS test collection project. There are two issues to constructing this corpus. One is a definition of domain that makes the terminology research different from IR keyword one. The other is a definition of lexical units in the texts that gives basic unit for terminology operations. In this paper, we discuss the approach to design the NACSIS corpus for solving these two issues, and show the status of NACSIS.
  • 吉岡真治  第9回設計工学 システム部門講演論文集  78  -81  1999
  • SEKIYA Takayuki, YOSHIOKA Masaharu, TOMIYAMA Tetsuo  Journal of the Japanese Society for Artificial Intelligence  14-  (6)  1051  -1060  1999  [Not refereed][Not invited]
     

    During design, a designer uses various computational tools, such as a geometric modeling system, analysis tools, and databases. To support these design processes, a system that can integrate such computational tools is required. For this purpose, we have been developing the Knowledge Intensive Engineering Framework (KIEF) system. KIEF enables the designer to build, evaluate, and modify design models on the system by providing him/her with a wide variety of engineering knowledge. A common ontology plays an important role in integrating tools and representing engineering knowledge. This paper describes the structure of the ontology in KIEF, and how it is used in a modeling processes on KIEF.

  • Yoshioka Masaharu  Research bulletin of the National Center for Science Information System  10-  (0)  73  -79  1998/03  [Not refereed][Not invited]
     
    Because of the finiteness of natural resources, environmental problems and so on, engineers are expected to evaluate their products from ever more various kinds of aspects. To support such claim, we proposed the concept of knowledge intensive engineering, in which various kinds of knowledge is used in a flexible and integrated manner in order to aim at generating more added-value. In addition, we also proposed Knowledge Intensive Engineering Framework (KIEF) system that forms a computational framework of knowledge intensive engineering. In this paper, I describe the concept of knowledge intensive engineering and the KIEF system. After that, I apply the KIEF system to architectural design to show the value of the system. To do so, I analyze knowledge that is used in architecture design and discuss how to implement the knowledge on the KIEF system. Finally, I show some results of this application and discuss the capability of the KIEF system.
  • 吉岡 真治  学術情報センター紀要  (10)  73  -79  1998/03  [Not refereed][Not invited]
  • 吉岡真治  1997年度冨山 梅田研究室論文集  1998
  • YOSHIOKA Masaharu, TOMIYAMA Tetsuo  Journal of the Japanese Society for Artificial Intelligence  13-  (2)  312  -319  1998  [Not refereed][Not invited]
     

    In this study, we propose a framework of an integrated modeling environment for design. Since a mechanical engineering design process requires various kinds of design object models, such as a geometric model, a kinematic model, and a finite element model, this environment should maintain consistency among these models. Furthermore, commercial design tools are practically used in design processes, and these must be integrated as well. For dealing with multiple design object models, we have formalized a concept called a metamodel. A metamodel is a model which represents relationships among concepts used in various design object models. This paper proposes a framework of the pluggable metamodel mechanism that allows to plug in existing design tools, to support a modeling on it, and to maintain consistency among them. We describe a prototype system and illustrate an example of ball screw design.

  • Y Umeda, M Ishii, M Yoshioka, Y Shimomura, T Tomiyama  AI EDAM-ARTIFICIAL INTELLIGENCE FOR ENGINEERING DESIGN ANALYSIS AND MANUFACTURING  10-  (4)  275  -288  1996/09  [Not refereed][Not invited]
     
    The relative significance of conceptual design to basic design or detail design is widely recognized, due to its influential roles in determining the product's fundamental features and development costs. Although there are some general methodologies dealing with functions in design, virtually no commercial CAD systems can support functional design, in particular so-called synthetic phase of design. Supporting the synthetic phase of conceptual design is one of the crucial issues of CAD systems with function modeling capabilities. In this paper, we propose a computer tool called a Function-Behavior-State (FBS) Modeler to support functional design not only in the analytical phase but also in the synthetic phase. To do so, the functional decomposition knowledge and physical features in the knowledge base of the modeler, and a subsystem Qualitative Process Abduction System (QPAS) play crucial roles. Modeling scheme of function in relation with behavior and structure and design process for conceptual design in the FBS Modeler are described. The advantages of the FBS Modeler are demonstrated by presenting two examples; namely, an experiment in which designers used this tool and the design of functionally redundant machines, which is a new design methodology for highly reliable machines, as its application.
  • 吉岡 真治, 冨山 哲男  精密工学会大会学術講演会講演論文集  1996-  (1)  5  -6  1996/03/01  [Not refereed][Not invited]
  • 吉岡真治  第14回設計シンポジウム講演論文集  40  -47  1996
  • 石井 理貴, 関谷 貴之, 吉岡 真治, 冨山 哲男  精密工学会大会学術講演会講演論文集  1995-  (2)  9  -10  1995/09/01  [Not refereed][Not invited]
  • Yoshioka Masaharu, Nakamura Masahiko, Tomiyama Tetsuo, Yoshikawa Hiroyuki  TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series C  61-  (581)  330  -337  1995  [Not refereed][Not invited]
     
    An intelligent CAD system aims at serving a designer as an intelligent, integrated design environment. For building such a system, development of a large scale knowledge base is necessary. To do so, design knowledge must be systematically formalized, made computable, and organized. Design knowledge has two categories ; i.e., design object knowledge and design process knowledge. In our previous reports, we have formalized a computable design process model to represent design process knowledge and the concept of a metamodel dealing with multiple design object models. This paper proposes a framework to achieve a harmonized combination of these two. In this framework, the computable design process model navigates the metamodel mechanism to generate and modify models of the design object. We describe a prototype system and illustrate an example of ship design.
  • YOSHIOKA Masaharu, ISHII Masaki, TOMIYAMA Tetsuo  人工知能学会全国大会論文集 = Proceedings of the Annual Conference of JSAI  8-  (0)  335  -338  1994/06/20  [Not refereed][Not invited]
  • ISHII Masaki, YOSHIOKA Masaharu, TOMIYAMA Tetsuo  人工知能学会全国大会論文集 = Proceedings of the Annual Conference of JSAI  8-  (0)  339  -342  1994/06/20  [Not refereed][Not invited]
  • 吉岡真治  1994年度精密工学会春季大会講演論文集  233  -234  1994
  • 吉岡 真治, 富山 哲男  東京大学工学部総合試験所年報  (52)  p177  -182  1993/09  [Not refereed][Not invited]
  • 吉岡真治, 石井理貴, 桐山孝司, 冨山哲男  人工知能学会知識ベースシステム研究会資料  26th-  1993

Awards & Honors

  • 2022/02 Linked Open Data チャレンジ 2021 実行委員会 オントロジー賞（オントロノミー合同会社）
     Wikipediaカテゴリオントロジー 
    受賞者: 北海道大学大学院情報科学研究院知識ベース研究室
  • 2022/02 Linked Open Data チャレンジ 2021 実行委員会 データ作成部門 優秀賞
     Wikipediaカテゴリオントロジー 
    受賞者: 北海道大学大学院情報科学研究院知識ベース研究室
  • 2015/06 人工知能学会 人工知能学会 研究会優秀賞
     複数国の新聞からの多観点比較による分析～GDELTデータを用いた分析～ 
    受賞者: 吉岡真治;神門典子

Research Grants & Projects

  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research
    Date (from‐to) : 2023/06 -2025/03 
    Author : 吉岡 真治, 原 真二郎, 長田 裕也
  • 日本学術振興会：科学研究費助成事業 挑戦的研究(萌芽)
    Date (from‐to) : 2021/07 -2023/03 
    Author : 吉岡 真治
     
    本研究では、特定の分野に興味を持つ研究者が収集する関係分野の論文を対象としたデータベースに基づいて、専門用語抽出などを行うことにより、用語間の共起関係に基づく分析や、時系列を考慮した研究動向分析を行うデータベースを基礎として、化学物質データベースなどのさまざまな科学技術データベースと連携させることにより、分野の研究者の研究活動を支援するシステムの構築を目指している。本年度は、これまでに作成した論文・図表データベースを様々な科学技術データベースと連携させるための手法について検討すると共に、その機能をこれまでに作成した論文・図表データベースに実装した。本データベースでは、PDFやRSS（Rich Site Summary:論文誌などが論文を発行する際に、abstractなどのメタデータを含む情報を配信する仕組み）を対象として、対象データの収集を行う。 本データベースでは、収集した各論文および、論文中に含まれる図表に対して、本文やキャプションから分類済みの用語抽出を行い、その出現頻度などをメタデータとして付与している。このメタデータを用いた検索を行うことにより、材料としてAを使っている論文（図表）で、利用されることが多いパラメータといった検索が可能となる。また、検索結果を、抽出した専門用語をハイライトすると共に、外部データベースへのリンクが埋め込まれた形で表示することで、科学技術データベースとの連携を実現している。また、論文からの情報抽出の研究として、論文からの化学反応情報の抽出に関する研究を進めると共に、化学反応データベースであるReaxysとの対応関係についての調査を行った。
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2019/04 -2022/03 
    Author : Utsuro Takehito
     
    In this project, we focused on Web pages explaining certain concepts, events, procedures etc. in a beginner friendly manner. In those Web pages, visual intelligibility in Web page layout and beginner friendly Web page texts are the requirements of Web contents for beginners. As a typical case study, this project focused on the factors of Web page visual intelligibility as well as beginner friendliness for explaining academic concepts. Then, this project proposed how to measure those factors as well as to identify those factors. Finally, in this project, we developed a system for recommending those Web pages explaining academic concepts based on those factors.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2018/04 -2022/03 
    Author : Yoshioka Masaharu
     
    In this research, we proposed a method to analyze news site stance of particular news topic using the distribution of polarity of the articles published by the news site from multiple perspectives. In this research, we confirmed that it is important to understand the interests on subtopics and/or genres of the news articles for detailed analysis of the news site stance. Based on this understanding we proposed a system that can analyze news site stance interactively by taking account the polarity of the related news articles for multiple subtopics and/or genres. We also proposed Wikipedia category ontology that is new knowledge resource extracted from Wikipedia category from the viewpoint of knowledge engineering. This new knowledge can be used for analyzing the type of the named entity.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research
    Date (from‐to) : 2019/06 -2021/03 
    Author : Yoshioka Masaharu
     
    Objective of this research project is proposal of an paper/figure database system with technical term management ecosystem for particular research group. This system constructs database based on the information (text, figure, caption, and metadata) extracted from the research papers collected by a research group. The system also supports to construct technical term dictionary management system and this dictionary is used for extracting technical terms from the texts. The system also provides a framework to analyze papers using these terms. We implement a system and construct database using research papers of interest and evaluate the effectiveness of the system. We plan to use this database for the ordinal research activities for long term evaluation.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (A)
    Date (from‐to) : 2016/04 -2020/03 
    Author : Kando Noriko
     
    The purpose of this study is to propose the interactive exploratory search technologies supporting users based on each user's situation. Firstly, we have conducted multiple laboratory-typed or crowdsourcing user studies and investigated various factors affect the search process and user's perceived satisfaction, such as the cognitive complexity of the search tasks. Secondly, we have investigated and tested the wide variety of tools and underlying technologies supporting user's exploratory such as inference of the user's situation based on logs and/or eye gaze, query recommendation, search results diversification, multi-facet search, and ostensive search. Thirdly the prototype of the ostensive search model-based interactive exploratory search and guide app for the museum was developed, using user's logs both in the virtual space on the app and the physical space in the museum, encouraged the unforgettable museum experience and introduced the context before and after the museum visits.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (C)
    Date (from‐to) : 2016/04 -2019/03 
    Author : Haraguchi Makoto
     
    When we read several stories or legal cases, we often find relational similarities between objects in spite of the difference of stories they belong to. The goal of this research is to design and implement an efficient algorithm for finding those structural similarities among more than two domains. For this purpose, we have designed such an algorithm making use of Non-Negative Matrix Factorization and Graph Mining techniques. The algorithm we have obtained is effective not only for text analysis but also for mining similar structure among databases as data warehouses.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2014/04 -2017/03 
    Author : UTSURO Takehito, NAKAGAWA Hiroshi
     
    In this project, we developed a system for comparative analysis of Web search information needs between languages. As for the source of collecting information in multiple languages, we focus on search engine suggests as well as question-answer sites in multiple languages. Given a specific query focus, we collect search information needs as well as Web pages and question-answer pairs for those search information needs. Then, developed a system for assisting the process of comparatively analyzing those collected search information needs and the specified Web pages / question-answer pairs between multiple languages and that of discovering differences of culture, concern, and opinions between multiple countries.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Challenging Exploratory Research
    Date (from‐to) : 2014/04 -2016/03 
    Author : Yoshioka Masaharu, HARA Shinjiroh
     
    In this project, we have constructed a corpus for extracting useful information related to the experiment of nanocrystal device development (e.g., Source material and parameters) and developed an automatic information extraction system NaDev that tries to extract such information from papers. By using NaDev, we can extract characteristic keywords related to the experiment (e.g., source material, experimental parameter) and these extracted keywords are used for calculating similarity among different papers and information retrieval system that have query construction support. In this project, we proposed a figure retrieval system that uses such parameter information as query.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2013/04 -2016/03 
    Author : Yoshioka Masaharu, KANDO Noriko
     
    In this research, we proposed a framework to extract named entity from the text with named entity type information by using DBPedia. We also proposed facet-biased topic model that take into account the type of named entity as facet. In addition, we implemented WC3 that analyzes consistency of metadata in articles that belongs to a particular Wikipedia category. We confirmed that WC3 can find out the articles that have inconsistent metadata and candidate articles for the category.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2011/04 -2014/03 
    Author : UTSURO TAKEHITO, YOSHIOKA Masaharu, INUI Takashi, NAKAGAWA Hiroshi, KIYOTA Yoji
     
    In this project, given a collection of multilingual documents available on the Web, we cross-lingually and cross-culturally compare facts and opinions that are observed in the collected documents. We especially focus on various facets including the following four of (i) to (iv), and realize a framework of assisting the process of discovering cross-cultural differences: (i) differences in specific issues and concerns within a topic, (ii) differences in time series features, (iii) differences in real life experiences, (iv) differences in opinions.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (A)
    Date (from‐to) : 2010/04 -2013/03 
    Author : OYAMA Keizo, AIZAWA Akiko, MIYAO Yusuke, SUN Yuan, KOBAYASHI Tetsuro, HAN Hao, KISHIDA Kazuaki, YAMANA Hayato, OKUMURA Manabu, YOSHIOKA Masaharu, ISHITA Emi, MURATA Tsuyoshi, EGUCHI Koji
     
    For understanding Web structure and users' behavior of information retrieval and browsing in an integrated way, and for extending it to various applications, we collected and introduced various data reflecting Web information structure and Web users' behavior (e.g. Web view log data, micro-blog data), obtained user data through questionnaire, and executed integrated analysis on them. Consequently, we obtained various findings through data such that there is a gap between information wanted to know and information wanted to inform, and that, through using Web portal sites, unexpected contact to various information occurs. Moreover, we proposed and studied various methods for advanced information access such as information recommendation and information retrieval based on the information obtained through the integrated analysis.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2009 -2011 
    Author : YOSHIOKA Masaharu, KANDO Noriko, SEKI Yohei
     
    In this research, we propose a news analysis system that supports to compare articles from news sites in different countries. This system can support to analyze articles from multi facet (e. g., opinion, named entity). In this research, we propose to construct named entity database and opinion analysis system. In addition, we also propose a framework to evaluate such interactive system by using information that characterizes change of the users' knowledge about task domain.
  • 日本学術振興会：科学研究費助成事業 特定領域研究
    Date (from‐to) : 2007 -2008 
    Author : 吉岡 真治, 原口 誠, 大久保 好章
     
    本研究では、情報抽出元である各々のWebサイトが持つ情報の偏りなどの特徴を分析する方法を提案すると共に、この特徴分析の結果を踏まえた意見集約・分析の方法を提案している。具体的には、全てのデータセットを含む全体のデータベースと特定の条件を満たす部分データベースにおけるデータの分布の違いに注目したコントラストセットマイニングを適用することにより、個々のサイトの特徴を分析する方法を提案する。 この目標に対して、本年度は、「コントラストセットマイニングによる個々のサイトの特徴分析」の考え方に基づくシステムを実装し、様々なレベルのユーザ実験を通した評価を行った。簡単にシステムを使用してもらってコメントをもらうというユーザ実験では、半数以上の利用者から、システムが有用であるというコメントをもらえた。一方、レポート作成課題の実施という本システムの利用目的に応じた課題の遂行という実験を行ったところ、異なる国のデータとの比較を行う際には、外国語の新聞を直接の比較対象とする必要性などが指摘された。この指摘を受け、機械翻訳システムと組み合わせることによりシステムを作成した。このシステムでは、ニュースなどの時事的なトピックに対する対訳辞書の不備を補うためのWikipediaを用いた対訳辞書の作成方法などの提案も行った。また、ピンポイントクラスタリングによるWebサイトの分析」の課題については、昨年度から継続して、フォーマルコンセプト(FCA)を用いた分枝限定法にもとづくトップN-FC探索手法の研究を行った。
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2007 -2008 
    Author : MAKOTO Haraguchi, ETSUJI Tomita, KIMIHITO Ito, MASAHARU Yoshioka, YOSHIAKI Okubo
     
    潜在的なクラスタ間接続を, たとえ, 外延的な重なりが小さな場合でも, 内包的な繋がりが高い場合は有意義なクラスタ間接続関係を与える「情間」と捉える.こうした「情間」は, クラスタの外延と内包に対する評価関数を設定することにより, 制約条件下の最適化問題として捉えることができ, 具体的には, 形式概念束探索における分枝限定アルゴリズムとして設計・開発・実証を行った.形式概念束は, 概念の抽象階層と見なすことができ, 主要なクラスタを繋ぐ適切な概念を抽象階層から見出すことを意味している.実証実験において, カテゴリとしては離れた異種文書群から, それらを跨ぐ, 機能的な概念を情間(クロスオーバ概念)として抽出することに成功している.
  • 日本学術振興会：科学研究費助成事業 若手研究(B)
    Date (from‐to) : 2005 -2007 
    Author : 吉岡 真治
     
    本研究では、問題解決を目的とした文書作成支援型知識マネージメントシステム(Documentation MAnagementsystem for Problem Solving:DMaPS)の構築を目的とした研究を実施した。このシステムでは、問題解決の支援、文書と形式化された知識の蓄積、文書からの知識の獲得というループの実現を目指している。この目的に対し、平成17年度は、問題解決の支援を行うために、ユーザが持っている問題に関する記述と類似している文書知識を提示することにより、問題解決の支援を行うシステムを作成した。本システムに対し、あらかじめ手作業により作成した知識を埋め込んだ文書群を与えることにより、問題解決の支援が可能であることを確認した。平成18年度以降は、主に、一般的な文書から問題解決に役立つ知識を獲得するための方法論に注目して研究を遂行した。具体的には、トラブル事例に関する知識獲得を目指して、新聞記事などから、ユーザが直面している問題に関連する訴訟の記事などを検索し、その検索結果から、関係する利害関係者とその関係を抽出するシステムの構築を行った。本システムでは、少数の訓練事例にたいして、複数の基準を用いた機械学習システムを適用し、その結果を組み合わせることによって、人手で作成したパターンでは抽出することができなかった多くの利害関係者の情報を獲得することが確認できた。また、文書分類の研究では、特許を関連する物理法則によって分類する考え方を提案し、Web上に存在する様々な物理法則に関する文書の情報などを利用して、具体的な特許文書を分類する方法を提案した。ただ、これらで得た知識は、DMaPSで利用可能なレベルの知識での形式化が困難であったため、今後は、形式的な知識作成に関する検討を進める必要がある。
  • 日本学術振興会：科学研究費助成事業 特定領域研究
    Date (from‐to) : 2004 -2005 
    Author : 吉岡 真治, 原口 誠, 大久保 好章
     
    本研究では、検索意図を推定し、それをユーザにも理解しやすい形式で表現できる情報検索システムを提案する。ここでは、検索者が用いる検索語の抽象度に注目して、検索意図の推定を行う。しかし、ユーザは、検索意図を表現するのに適切な抽象度の概念を必ずしも用いない場合がある。そのため、検索意図に応じた適切な抽象度の概念を検索キーワードに用いると、ユーザにも理解しやすく効率的な検索キーワードになると考えられる。 そのため、本研究では、電子化辞書やシソーラスに記述されている概念階層構造を利用し、検索意図に応じて検索キーワードを適合的に汎化し、新たな検索式を作成する情報検索システムを提案する。ここで、適切な抽象度の汎化とは、検索キーワードが持つ正解判定の分別能力に関する情報を多く保存する汎化の事である。 本年度は、検索式における検索語と検索要求のミスマッチという観点から検索課題の持つ性質の分類を行うと共に、検索語拡張に与える影響を分析した。その結果、検索語と検索要求のミスマッチのタイプに応じて、検索語拡張に求められる機能も異なり、単一の検索語拡張手法では、検索性能を改善させる場合だけでなく、性能を悪化させる場合があることが確認された。よって、検索語拡張の洗練化を行うためには、検索語と検索要求のミスマッチに応じた検索語拡張の方法の提案と、ミスマッチのタイプの推定を行うことが必須であることを提案した。
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2004 -2005 
    Author : HARAGUCHI Makoto, TANAKA Yuzuru, KAKUTA Tokuyasu, YOSHIOKA Masaharu, OKUBO Yoshiaki
     
    In the research project, an algorithm for extracting common abstract event sequences, given two or more documents, is presented. In order to avoid combinatorial explosions in matching more than two documents, the algorithm consists of two phases. The first phase is basically a text summarization system taking balances between the importance of sentences in each segment in a given document and the importance of sentences to connect several segments. The latter importance is used to extract contextual sentences involving contextual words. In order to separate the notion of importance into the two as in the above, we compute a chunk of core sentences in each segment by a clique finding algorithm, and then calculate the degree of latter importance of sentences from the chunk just in the way used in KeyGraph. Finally, the overall importance is determined by a scheme very similar to topic-sensitive PageRank. We have made some experiments for newspaper articles, and verified its effectiveness. In the second phase, we use the summarized document obtained in the first phase as a kind of source document. The summarized document preserves the structure of the original document at more abstract level. Therefore, for each event in the source, it suffices to find a similar event in a given target document. This drastically reduces the computational complexity need to build correspondence between the two documents. As a result, several document set is now summarized from a viewpoint of the source document by means of analogy.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 2002 -2004 
    Author : TAKEDA Hideaki, SHIMOMURA Yoshiki, YOSHIOKA Masaharu
     
    This project presents "Universal Abduction Studio" (UAS) that is a new approach to computational support for designers and discusses technologies to implement it. Although design support systems have been well developed for geometric and detail design stages, they are still unsuccessful in the conceptual design stage. In our opinion, the main difficulty comes from incomplete and insufficient understanding about design knowledge and its operations that play a crucial role in conceptual design. In contrast, however, recently thanks to development of the Internet technologies, more and more knowledge is accumulated and available electronically. It then becomes an interesting research question how to apply such an enormous amount of diverse knowledge to conceptual design. UAS is not intended to be an automated design system. UAS is an environment in which designers can manage and apply knowledge to arrive at new, creative design collaboratively with the system. UAS provides the designer with knowledge and reasoning functions, while the designer selects operations and evaluates design results iteratively. we show a novel knowledge representation to support creative design process. Our RDF-based representation consists of modules for knowledge repositories like predicate and object ontologies. A user can compose an individual model for knowledge by selecting and combining suitable elements in these knowledge repositories. This representation solves two characteristics observed in our case study analysis, ie., multiplicity of interpretation and variety of knowledge. The graph structure can generate multiple interpretations with minimum modification such as redirecting "if" and "then" links. Furthermore, since we separate metamodel and aspect-specific models, we can obtain various aspect-specific models from a single metamodel. The problem to handle with variety of knowledge is balance of expansivity and consistency. Since we provide categories of knowledge like object and predicate ontologies in order to clarify roles of knowledge, we can add easily new knowledge like domain-specific knowledge according to its roles without avoiding messy.
  • 日本学術振興会：科学研究費助成事業 特定領域研究
    Date (from‐to) : 2003 -2003 
    Author : 吉岡 真治, 大久保 好章, 原口 誠
     
    本研究では、検索意図を推定し、それをユーザにも理解しやすい形式で表現できる情報検索システムを提案する。ここでは、検索者が用いる検索語の抽象度に注目して、検索意図の推定を行う。しかし、ユーザは、検索意図を表現するのに適切な抽象度の概念を必ずしも用いない場合がある。そのため、検索意図に応じた適切な抽象度の概念を検索キーワードに用いると、ユーザにも理解しやすく効率的な検索キーワードになると考えられる。 そのため、本研究では、電子化辞書やシソーラスに記述されている概念階層構造を利用し、検索意図に応じて検索キーワードを適合的に汎化し、新たな検索式を作成する情報検索システムを提案する。ここで、適切な抽象度の汎化とは、検索キーワードが持つ正解判定の分別能力に関する情報を多く保存する汎化の事である。 本年度は、この適合的汎化による検索拡張が、正解文書において補完関係にある検索語の発見を目指しているという点に注目し、適合的汎化による検索拡張の結果をBoolean式として表現する方法について検討を行った。さらに、実際にユーザが利用可能なインターフェースの作成を行った。また、文書中の語の役割を反映した抽象概念の提案手法についても検討を行った。これらの手法の有効性は、実際のユーザ実験によって検討する必要性がある。次年度以降は、ユーザ実験などを通して、本手法の有効性の検証と手法の洗練化を図っていきたいと考えている。
  • 日本学術振興会：科学研究費助成事業 特定領域研究
    Date (from‐to) : 2002 -2002 
    Author : 吉岡 真治, 大久保 好章, 原口 誠
     
    本研究では、検索者にも理解しやすい概念階層の汎化という考え方を利用して、ユーザの検索意図を明示化すると共に、精度落ちを抑えた情報検索システムを提案している。この概念階層を利用する情報検索システムでは、汎化の対象としてどの様な語を選ぶのかという問題に加え、あらかじめ用意する辞書における概念階層の性質に結果が依存するという問題がある。 本年度は、汎化のための方法論として、適合文書群に共通する概念を利用する方法を検討した。学術論文のアブストラクトを用いたNTCIRや新聞記事を用いたIREXの2つの情報検索のテストコレクションを利用し提案手法のシステムの検索性能への影響を検討した。その結果、提案手法により検索性能が向上することが確認できた。 次に、辞書の性質を議論するために、昨年度までの研究で用いていたEDR電子化辞書に加え、日本語概念語彙体系という異なるタイプのシソーラスを用いた場合の影響に関する分析を行った。検索実験の結果、日本語概念語彙体系を用いた場合に、与えた適合文書へのオーバーフィットの度合いが高いことが確認された。これは、日本語概念語彙体系とEDRが持つ概念カテゴリーの粒度の違いによるものであると考えられる。日本語概念語彙体系とEDRが持つ概念カテゴリーの数を単純に比較すると、EDRは約50000個(専門用語辞書のものを含む)であり、日本語概念語彙体系は約3000である。よって、適合的汎化による情報検索システムで用いるシソーラスは、ある程度、粒度の細かい概念を扱うことができるシソーラスであることが望まれることが確認された。
  • 日本学術振興会：科学研究費助成事業 若手研究(B)
    Date (from‐to) : 2001 -2002 
    Author : 吉岡 真治
     
    本研究では、設計問題に代表される問題解決の過程を問題を表現するためのモデルとそこで用いられた知識さらに、その際に記述された文書を組み合わせて保存することにより、設計文書の体系的管理と再利用を実現することを目的とした。その成果として、問題解決のための文書管理システムDMaPS(Documentation Management system for Problem Solving)を提案した。 このDMaPSでは、ひとつのサブ問題の問題解決の過程を単位として考えた文書知識を利用して、文書の管理と問題解決の支援を行う。この文書知識では、文書を単純なテキスト情報として保存するだけではなく、知識を適用する背景となる条件や形式的な知識を一階述語論理によって表現する、この二つの表現を組み合わせることにより、マニュアルなどの形式知とノウハウのように形式化が困難な情報を組み合わせて取り扱うことが可能となる。 本システムでは、この文書知識を用いることにより、ユーザの問題解決の支援を行う。具体的には、ユーザが直面している問題との一階述語論理による形式的表現に関する類似性とキーワードによる情報検索による類似性という異なる観点から知識の検索を行うことができ、より柔軟な知識の検索を実現する。また、検索された知識を取り込むことにより、問題解決の過程を説明するための文書が半自動的に作成でき、ユーザは個別事例に適応させる部分の記述に専念することができる。さらに、本システムでは、知識のバージョン管理の機構を取り込んでおり、各文書問題解決の際に利用した知識のバージョンをチェックすることにより、既存の設計文書をそのまま再利用してよいのか、再検討が必要なのかを調べることが可能になっている。 本研究では、建築構造物の基礎設計を例にとり、知識記述の実験を行い、体系的な知識整理が可能であることを示した。
  • 日本学術振興会：科学研究費助成事業 特定領域研究(C)
    Date (from‐to) : 2001 -2001 
    Author : 吉岡 真治, 大久保 好章, 原口 誠
     
    本研究では、検索意図を推定し、それをユーザにも理解しやすい形式で表現できる情報検索システムを提案する。ここでは、検索者が用いる検索語の抽象度に注目して、検索意図の推定を行う。 しかし、ユーザは、検索意図を表現するのに適切な抽象度の概念を必ずしも用いない場合がある。そのため、検索意図に応じた適切な抽象度の概念を検索キーワードに用いると、ユーザにも理解しやすく効率的な検索キーワードになると考えられる。 そのため、本研究では、電子化辞書やシソーラスに記述されている概念階層構造を利用し、検索意図に応じて検索キーワードを適合的に汎化する情報検索システムを提案する。ここで、適切な抽象度の汎化とは、検索キーワードが持つ正解判定の分別能力に関する情報を多く保存する汎化の事である。 一方、電子化辞書やシソーラスでは、一般的な概念階層が記述されているために、多段階の階層を用いた汎化を行った際に、一度に大量の語を代表する概念に汎化されてしまう問題がある。例えば、「ビデオ」と「DVD」を汎化した概念として「映像装置」という概念を考えたときに、8mmカメラは含むことは、検索意図と必ずしも一致しないと考えられる。そのため、ユーザが持つ細かな検索意図を適切に表現するために、検索目的に応じた適切な抽象概念を設定し、概念階層を再構築することが必要である。 本年度は、概念の適切な抽象化のための指標として、キーワードと正解文書の相互情報量に基づく関連性の指標を利用し、情報検索システムを構築した。本システムは、5件の正解文書を与えることにより、検索語の汎化を行い、検索者の意図が理解しやすい検索式が作成可能である。また、この検索式による検索は、他の検索システムと比較して十分な検索精度を持っていることを確認した。
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research (B)
    Date (from‐to) : 1999 -2001 
    Author : TAKEDA Hideaki, TAKEUCHI Koichi, UENO Atsushi, NISHIDA Toyoaki
     
    In this research, we proposed concept "knowledge as media" to realize knowledge lifecycle for communication support on the network. In the traditional view of artificial intelligence, Knowledge exists within agents such as human and computer programs. On the contrary, we regard knowledge exists among agents, i. e., knowledge is media which enables communication among agents. Under this view, we investigated knowledge from the abstract way to the practical one. Ontology is the most abstract form of knowledge, knowledge for community and knowledge for environments are intermediate, and the embodied knowledge is the most practical one. We have three major results. One is community-support system called TelMeA in which interface agents as participants can form a field of communication. We tested this system by the psychological experiment and showed that using interface agents like this system is effective for communication. The second result is a system to capture personal knowledge. The system called MindHeap can gather knowledge through WWWW browsing. If a user can pick up sentences in WWWW, the system memorizes the sentences and their contexts. The sentences are connected to each other through their contexts in this system. The user can retrieve related memorized sentences automatically and even summarize WWW pages based on memorized information. The third result is embodiment-based learning, in which embodiment knowledge is captured through actions. The agent can learn how its body is by moving around and sensing the environment.
  • Japan Society for the Promotion of Science：Grants-in-Aid for Scientific Research Grant-in-Aid for international Scientific Research
    Date (from‐to) : 1995 -1997 
    Author : INOSE Hiroshi, SAKAI Yoshinori, SEZAKI Kaoru, IIDA Noriko, YASUDA Yasuhiko, FUJINO Takayuki, YOSHIOKA Shinji, ZHAO Ihei, ABE Shuji, UOSE Hisao, LANDWEBER Lawrence, AIZAWA Akiko, KEI Usei, OHYAMA Keizo, MATSUKATA Jun, HASHIZUME Hiromichi, ADACHI Jun, MIYAZAWA Akira, ASANO Shoichiro, HAMADA Takashi, ONO Kinji
     
    In recent years, research networks have extremely diversified, not only in developed coutries but also in newly developing contries and region, including mid and east europe, mid and south america, africa and asian coutries. As the result, international coordination among research networks is of important to establish international communication links. The CCIRN (Coordination Committee for Intercontinental Research Networks) plays primary role of such coordination. In addition to this, need for braodband communiation links is emphasized in order ato stimulate higher end researches. In this context, the CCIRN is now discussing how to interwork such broadband communication links and networks. United States had lauched new interconnecting point among broadband networks in Chikago, IL., which is called STAR-TAP (Science, Technology and Research, Transit Access Point). Investigators joined to the study have made international cooperation activities during three years, and finally made success to interconnected japanese boradbnad network (SINET) to STAR-TAP in October, 1997. European Commuinity has also put in operation of braodband research network (TEM-34). SINET has been cooperating to european resarch networks for three years, and consequently SINET has changed its connection point to TEM-34 PoP (Point of Presence) to London, UK.Circuit to London will be in operation from early April, 1998. That is second result of the study. Beside the actual interconnection, investigators have been contributing international standardization at IETF on ATM based networking. Several proposals which are now available as Internet Drafts, including ATM address resolution scheme, ATM inteconnection control information and high-performance protocol sepcifications, are proposed from the study. These activities are evaluated as future basis for international connection of research networks.
  • 大規模知識ベースを用いた設計支援の研究
    Date (from‐to) : 1996

Educational Activities

Committee Membership

  • 2009 - Today   Steering committee member of Asian Information Retrieval Societies
  • 2017/12 -2021/06   Tokyo ACM SIGIR Chapter   Treasurer

Copyright © MEDIA FUSION Co.,Ltd. All rights reserved.