査読あり, その他, NICOGRAPH 2021, フルペーパー,2021, 学生奨励賞, 花紋スモッキングの展開図に対するモジュール, 夛田美彩; 吉田哲也, 2021年11月
査読あり, その他, 情報処理学会論文誌:数理モデル化と応用, リジッドルームのための織物組織図の近似最適化, 吉田哲也, 2021年08月, 14, 3, 76, 84, 研究論文(学術雑誌)
査読あり, その他, 芸術科学会論文誌, 花紋スモッキングの展開図に基づく紐の生成, 吉田哲也; 藤田真奈美, 2021年03月, 20, 1, 10, 20
査読あり, その他, 芸術科学会論文誌, 正多角形の貼り合わせを用いた花紋スモッキングの 組み合わせの拡張, 夛田美沙; 吉田哲也, 2020年11月, 19, 4, 40, 48, 研究論文(学術雑誌)
査読あり, その他, 芸術科学会論文誌, 花紋折りに基づくスモッキングのパターン作成と組 み合わせのデザイン, 吉田哲也,藤﨑千晶, 2020年06月, 19, 2, 9, 24, 研究論文(学術雑誌)
査読あり, その他, 日本建築学会計画系論文集, ネパー ルの世界文化遺産登録都市における都市型住居の外観意匠類型 – バクタプ ル東部の都市街区を事例に–, 飛鳥濱岡; 山本直彦; 吉田哲也; 宮内杏里; 増井正哉; 向井洋一, 2019年01月, 84, 756, 425, 435, 研究論文(学術雑誌)
査読あり, 英語, COMPUTATIONAL INTELLIGENCE, COMMUNITY STRUCTURE-BASED APPROACH FOR NETWORK IMMUNIZATION, Tetsuya Yoshida; Yuu Yamada, We propose a community structure-based approach that does not require community labels of nodes, for network immunization. Social networks have been widely used as daily communication infrastructures these days. However, fast spreading of information over networks may have downsides such as computer viruses or epidemics of diseases. Because contamination is propagated among subgraphs (communities) along links in a network, use of community structure of the network would be effective for network immunization. However, despite various research efforts, it is still difficult to identify ground-truth community labels of nodes in a network. Because communities are often interwoven through intermediate nodes, we propose to identify such nodes based on the community structure of a network without requiring community labels. By regarding the community structure in terms of nodes, we construct a vector representation of nodes based on a quality measure of communities. The distribution of the constructed vectors is used for immunizing intermediate nodes among communities, through the hybrid use of the norm and the relation in the vector representation. Experiments are conducted over both synthetic and real-world networks, and our approach is compared with other network centrality-based approaches. The results are encouraging and indicate that it is worth pursuing this path., 2017年02月, 33, 1, 77, 98, 研究論文(学術雑誌), 10.1111/coin.12082
査読あり, その他, Computational Intelligence, A Graph-based approach for Semi-Supervised Clustering, Yoshida, T, 2014年05月, 30, 2, 263, 284, 研究論文(学術雑誌)
査読あり, その他, Social Network Analysis and Mining, Weighted line graphs for overlapping community discovery, Tetsuya Yoshida, 2013年12月, 3, 4, 1001, 1013, 研究論文(学術雑誌), 10.1007/s13278-013-0104-1
査読あり, その他, International Journal of Knowledge-Based & Intelligent Engineering Systems, Rectifying the representation learned by Non-negative Matrix Factorization, Yoshida, T, 2013年11月, 17, 4, 279, 290, 研究論文(学術雑誌)
査読あり, その他, Journal of Intelligent Information Systems, Toward finding hidden communities based on user profile, Tetsuya Yoshida, 2013年04月, 40, 2, 189, 209, 研究論文(学術雑誌), 10.1007/s10844-011-0175-2
査読あり, 英語, ACTIVE MEDIA TECHNOLOGY, AMT 2013, Toward Robust and Fast Two-Dimensional Linear Discriminant Analysis, Tetsuya Yoshida; Yuu Yamada, This paper presents an approach toward robust and fast Two-Dimensional Linear Discriminant Analysis (2DLDA). 2DLDA is an extension of Linear Discriminant Analysis (LDA) for 2-dimensional objects such as images. Linear transformation matrices are iteratively calculated based on the eigenvectors of asymmetric matrices in 2DLDA. However, repeated calculation of eigenvectors of asymmetric matrices may lead to unstable performance. We propose to use simultaneous diagonalization of scatter matrices so that eigenvectors can be stably calculated. Furthermore, for fast calculation, we propose to use approximate decomposition of a scatter matrix based on its several leading eigenvectors. Preliminary experiments are conducted to investigate the effectiveness of our approach. Results are encouraging, and indicate that our approach can achieve comparative performance with the original 2DLDA with reduced computation time., 2013年, 8210, 126, 135, 研究論文(国際会議プロシーディングス)
査読あり, 日本語, 情報処理学会論文誌:数理モデル化と応用,, 重複コミュニティ発見のための重み付き線グラフ, 吉田 哲也, 2012年07月, 5, 3, 79, 88, 研究論文(学術雑誌)
査読あり, 日本語, 情報処理学会論文誌:数理モデル化と応用,, ネットワークのノード情報を考慮した正則化モジュラリティ固有空 間法, 吉田 哲也, 2012年07月, 5, 1, 65, 73, 研究論文(学術雑誌)
査読あり, 英語, INTELLIGENT DECISION TECHNOLOGIES (IDT'2012), VOL 2, Immunization of Networks via Modularity Based Node Representation, Tetsuya Yoshida; Yuu Yamada, We propose an approach for immunization of networks via modularity based node representation. Immunization of networks has often been conducted by removing nodes with large centrality so that the whole network can be fragmented into smaller subgraphs. Since contamination is propagated among subgraphs (communities) along links in a network, besides centrality, utilization of community structure seems effective for immunization. However, despite various efforts, it is still difficult to identify true community labels in a network. Toward effective immunization of networks, we propose to remove nodes between communities without identifying community labels of nodes. By exploiting the vector representation of nodes based on the modularity matrix of a network, we propose to utilize not only the norm of vectors, but also the relation among vectors. Two heuristic scoring functions are proposed based on the inner products of vector representation and their filtering in terms of vector angle. Preliminary experiments are conducted over synthetic networks and real-world networks, and compared with other centrality based immunization strategies., 2012年, 16, 33, 44, 研究論文(国際会議プロシーディングス), 10.1007/978-3-642-29920-9_4
査読あり, 英語, INTELLIGENT DECISION TECHNOLOGIES (IDT'2012), VOL 2, Line Graph for Weighted Networks toward Overlapping Community Discovery, Tetsuya Yoshida, We propose generalized line graph for weighted networks toward overlapping community discovery from the networks. Community discovery from a network has often been conducted by assigning each node in a network only to one community. However, in real world networks, a node (e.g., user) might belong to several communities. For undirected networks without self-loops, we propose to generalize line graph by defining the weights in the line graph based on the weights in the original network. Based on the line graph representation, a node can be assigned to more than one community by assigning the links adjacent to the node to the corresponding communities. Various properties of the proposed generalized line graph are clarified, and the properties indicate that our proposal is a natural extension of the conventional line graph. Preliminary experiments are conducted over several real-world networks, and the results indicate that the proposed generalized line graph can improve the quality of the discovered overlapping communities., 2012年, 16, 403, 413, 研究論文(国際会議プロシーディングス)
査読あり, 英語, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), A comparative study of community structure based node scores for network immunization, Yuu Yamada; Tetsuya Yoshida, Network immunization has often been conducted by removing nodes with large network centrality so that the whole network can be fragmented into smaller subgraphs. Since contamination (e.g., virus) is propagated among subgraphs (communities) along links in a network, besides centrality, utilization of community structure seems effective for immunization. We have proposed community structure based node scores in terms of a vector representation of nodes in a network. In this paper we report a comparative study of our node scores over both synthetic and real-world networks. The characteristics of the node scores are clarified through the visualization of networks. Extensive experiments are conducted to compare the node scores with other centrality based immunization strategies. The results are encouraging and indicate that the node scores are promising. © 2012 Springer-Verlag., 2012年, 7669, 328, 337, 研究論文(国際会議プロシーディングス), 10.1007/978-3-642-35236-2_33
査読あり, 英語, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Influence of erroneous pairwise constraints in semi-supervised clustering, Tetsuya Yoshida, Side information such as pairwise constraints is useful to improve the clustering performance in general. However, constraints are not always error free in general. When erroneous constraints are specified as side information, treating them as hard constraints could have the disadvantage since strengthening incorrect or erroneous constraints can lead to performance degradation. In this paper we conduct extensive experiments to investigate the influence of erroneous pairwise constraints over various document datasets. Several state-of-the-art semi-supervised clustering methods with graph representation were evaluated with respect to the type of constraints as well as the number of constraints. Experimental results confirmed that treating pairwise constraints as hard constraints is vulnerable to erroneous ones. However, the results also revealed that the influence of erroneous constraints depends on how the constraints are exploited inside a learning algorithm. © 2012 Springer-Verlag., 2012年, 7669, 43, 52, 研究論文(国際会議プロシーディングス), 10.1007/978-3-642-35236-2_5
査読あり, 英語, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Community structure based node scores for network immunization, Tetsuya Yoshida; Yuu Yamada, We propose community structure based node scores for network immunization. Since epidemics (e.g, virus) are propagated among groups of nodes (communities) in a network, network immunization has often been conducted by removing nodes with large score (e.g., centrality) so that the major part of the network can be protected from the contamination. Since communities are often interwoven through intermediating nodes, we propose to identify such nodes based on the community structure of a network. By regarding the community structure in terms of nodes, we construct a vector representation of each node based on a quality measure of communities for node partitioning. Two types of node score are proposed based on the direction and the norm of the constructed node vectors. © 2012 Springer-Verlag., 2012年, 7458, 899, 902, 研究論文(国際会議プロシーディングス), 10.1007/978-3-642-32695-0_95
査読あり, 英語, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Overlapping community discovery via weighted line graphs of networks, Tetsuya Yoshida, We propose an approach for overlapping community discovery via weighted line graphs of networks. For undirected connected networks without self-loops, we generalize previous weighted line graphs by: 1) defining weights of a line graph based on the weights in the original network, and 2) removing self-loops in weighted line graphs, while sustaining their properties. By applying some off-the-shelf node partitioning method to the weighted line graph, a node in the original network can be assigned to more than one community based on the community labels of its adjacent links. Various properties of the proposed weighted line graphs are clarified. Furthermore, we propose a generalized quality measure for soft assignment of nodes in overlapping communities. © 2012 Springer-Verlag., 2012年, 7458, 895, 898, 研究論文(国際会議プロシーディングス), 10.1007/978-3-642-32695-0_94
査読あり, 英語, International Journal of Knowledge-Based and Intelligent Engineering Systems, A re-coloring approach for graph b-coloring based clustering, Tetsuya Yoshida; Hiroki Ogino, This paper proposes a re-coloring approach for graph b-coloring based clustering. Based on the notion of graph b-coloring in graph theory, a b-coloring based clustering method was proposed. However, previous method did not explicitly consider the quality of clusters, and could not find out better clusters which satisfy the properties of b-coloring. Although a greedy re-coloring algorithm was proposed to reflect the quality of clusters, it was still restrictive in terms of the explored search space due to its greedy and sequential re-coloring process. We aim at overcoming the limitations by enlarging the search space for re-coloring, while guaranteeing the b-coloring properties. In our approach, the vertices in a graph are divided into two disjoint subsets based on the properties of b-coloring. A best first re-coloring algorithm is proposed to realize non-greedy search for the admissible colors of vertices. A color exchange algorithm is proposed to remedy the problem in sequential re-coloring. These algorithms are orthogonal to each other with respect to the re-colored vertices, and thus can be utilized in conjunction. The proposed approach was evaluated against several UCI benchmark datasets. The results are encouraging and indicate the effectiveness of the proposed method, especially with respect to the ground truth micro-averaged precision. © 2012 - IOS Press and the authors. All rights reserved., 2012年, 16, 2, 117, 128, 研究論文(学術雑誌), 10.3233/KES-2011-0236
査読あり, 日本語, 情報処理学 会論文誌:数理モデル化と応用, 特徴表現のスパース性を考慮したNMF, 木村 圭吾; 吉田 哲也, 2012年, 5, 1, 21, 29, 研究論文(学術雑誌)
査読あり, その他, Journal of Intelligent Information Systems, A graph model for mutual information based clustering, Tetsuya Yoshida, 2011年10月, 37, 2, 187, 216, 研究論文(学術雑誌), 10.1007/s10844-010-0132-5
査読あり, その他, NMFを用いた表現学習に対するコレスキー分解を用い た補正法, 吉田哲也; 荻野広樹, 2011年07月, 4, 3, 94, 101
日本語, 人工知能学会論文誌 = Transactions of the Japanese Society for Artificial Intelligence : AI, The Japanese Society for Artificial Intelligence, 記述長に基づく適応的 Ripple Down Rules 法, 吉田 哲也; 和田 卓也; 元田 浩; 鷲尾 隆, A knowledge acquisition method Ripple Down Rules (RDR) can directly acquire and encode knowledge from human experts. It is an incremental acquisition method and each new piece of knowledge is added as an exception to the existing knowledge base. Past researches on RDR method assume that the problem domain is stable. This is not the case in reality, especially when an environment changes. Things change over time. This paper proposes an adaptive Ripple Down Rules method based on the Minimum Description Length Principle aiming at knowledge acquisition in a dynamically changing environment. We consider the change in the correspondence between attribute-values and class labels as a typical change in the environment. When such a change occurs, some pieces of knowledge previously acquired become worthless, and the existence of such knowledge may hinder acquisition of new knowledge. In our approach knowledge deletion is carried out as well as knowledge acquisition so that useless knowledge is properly discarded to ensure efficient knowledge acquisition while maintaining the prediction accuracy for future data. Furthermore, pruning is incorporated into the incremental knowledge acquisition in RDR to improve the prediction accuracy of the constructed knowledge base. Experiments were conducted by simulating the change in the correspondence between attribute-values and class labels using the datasets in UCI repository. The results are encouraging., 2004年11月01日, 19, 460, 471, 10.1527/tjsai.19.460
査読あり, 日本語, 人工知能学会論文誌, AGMアルゴリズムの高速化と立体構造解析への適用, Yoshio Nishimura; Takashi Washio; Tetsuya Yoshida; Hiroshi Motoda; Akihiro Inokuchi; Takashi Okada, Apriori-based Graph Mining (AGM) algorithm efficiently extracts all the subgraph patterns which frequently appear in graph structured data. The algorithm can deal with general graph structured data with multiple labels of vartices and edges, and is capable of analyzing the topological structure of graphs. In this paper, we propose a new method to analyze graph structured data for a 3-dimensional coordinate by AGM. In this method the distance between each vertex of a graph is calculated and added to the edge label so that AGM can handle 3-dimensional graph structured data. One problem in our approach is that the number of edge labels increases, which results in the increase of computational time to extract subgraph patterns. To alleviate this problem, we also propose a faster algorithm of AGM by adding an extra constraint to reduce the number of generated candidates for seeking frequent subgraphs. Chemical compounds with dopamine antagonist in MDDR database were analyzed by AGM to characterize their 3-dimensional chemical structure and correlation with physiological activity., 2003年, 18, 5, 257, 268, 研究論文(学術雑誌), 10.1527/tjsai.18.257
査読あり, 英語, 電気学会論文誌C, 多様な構造化に基づいた概念相違検出手法, Teruyuki Kondo; Shogo Nishida; Tetsuya Yoshida, When people carry out collaborative works, conceptual difference due to different backgrounds and knowledge can hinder their communication and deteriorate collaboration. We have been carrying out research on detecting conceptual difference by focusing on the situation in which different symbols are used to denote the same meaning and the same symbols are used to denote different meaning. In our approach each user's knowledge is represented as a decision tree respectively and difference in concept is detected as difference in the structure of decision trees. This paper points out some problems in constructing a unique decision tree based on a single information criterion. Based on the idea of diverse structure, this paper proposes a new method for increasing the performance of detection by constructing multiple decision trees with diverse structure. Genetic algorithm is utilized to realize the idea of diverse structure and experiments with motor diagnosis cases on the implemented system confirmed the i provement on the performance of detection. © 2003, The Institute of Electrical Engineers of Japan. All rights reserved., 2003年, 123, 2, 345, 354, 研究論文(学術雑誌), 10.1541/ieejeiss.123.345
査読無し, 日本語, 電子情報通信学会論文誌. D-I, 情報・システム, I-情報処理, 一般社団法人電子情報通信学会, 決定木の相関関係に基づいた概念相違検出手法, 大西 健介; 吉田 哲也; 西田 正吾, 複数の人間が共同作業を行う場合には,背景知識や専門領域の違いから派生する概念の相違が意思疎通を阻害する要因となる.筆者らは従来から概念の相違として異なるシンボルを同じ意味に用いたり,同じシンボルを異なる意味に用いたりする場合を考え,ユーザの知識を決定木として表現し,概念の相違を決定木の構造の相違としてとらえることでその検出を行うことに取り組んできた.本論文では,従来の各ユーザの知識から仮想的な知識を生成して決定木を構築することにより相違を検出する手法における問題点を指摘し,その解決策として決定木におけるクラス分類の行われ方からボトムアップに決定木間の相関関係を構築して相違検出を行う手法を提案する.提案した手法をプロトタイプに実装し,モータの故障診断事例に対する実験を通じて検出精度の向上を確認した., 2002年07月25日, 85, 8, 784, 797
査読あり, 日本語, 電子情報通信学会論文誌D-I, 決定木の相関関係に基づいた概念相違検出手法, 2002年, vol.J85, No.8, pp.784-797, 研究論文(学術雑誌)
査読あり, 日本語, ヒューマンインタフェース学会論文誌, ヒュ-マンインタフェ-ス学会, 検索キーワード間の関連性を用いた情報獲得支援システム, 新開 大樹; 吉田 哲也; 西田 正吾, 2002年, Vol.4, No.4,pp.19-30, 4, 207, 217, 研究論文(学術雑誌)
査読あり, 日本語, ヒューマンインタフェース学会論文誌, ヒュ-マンインタフェ-ス学会, イメージに基づいたWebページデザイン支援システム, 渡辺 正人; 吉田 哲也; 才脇 直樹; 土方 嘉徳; 西田 正吾, 2001年, Vol.3, No.4,pp.287-297, 4, 73, 83, 研究論文(学術雑誌)
査読あり, 英語, ADVANCES IN WEB-AGE INFORMATION MANAGEMENT, PROCEEDINGS, Utilizing the correlation between query keywords for information retrieval, T Yoshida; D Shinkai; S Nishida, This paper proposes a method to utilize the correlation between query keywords in information retrieval toward extrapolating their semantic information. In accordance with the rapid development of internet and WWW (World Wide Web), it has been getting more and more hard to pinpoint the appropriate document from the huge information resource. Various search engines have been developed to retrieve the appropriate information based on the keywords, however, it is hard for the user to specify the keywords enough to pinpoint the appropriate documents. Since there often exist some semantic correlation between the keywords, this paper proposes to finding out the another keyword by utilizing their semantic correlation in order to narrow the scope of seach. Experiments were carried out to investigate the effectiveness of the proposed method and the result hinted the effectiveness of our approach as a pre-processing to narrow the scope of search for search engines., 2001年, 2118, 49, 59, 研究論文(学術雑誌)
査読無し, 日本語, 電気学会論文誌. C, 電子・情報・システム部門誌 = The transactions of the Institute of Electrical Engineers of Japan. C, A publication of Electronics, Information and System Society, 電気学会, 情報提供者によるユーザのナビゲーションを支援する適応型ハイパーメディアシステム, 土方 嘉徳; 吉田 哲也; 西田 正吾, 2000年11月01日, 120, 11, 1720, 1731
査読あり, 日本語, ヒューマンインタフェース学会論文誌, ユーザ参加型設計のそれの事例を利用した意図理解支援インタフェース, 2000年, 2, 2, 87, 95, 研究論文(学術雑誌)
査読あり, 英語, IEEE RO-MAN 2000: 9TH IEEE INTERNATIONAL WORKSHOP ON ROBOT AND HUMAN INTERACTIVE COMMUNICATION, PROCEEDINGS, A method for supporting software design based on comment on software, T Yoshida; K Hashimoto; T Yamaoka; S Nishida, This paper proposes a method for supporting software design by utilizing the comment which is usually put on software. It is believed that the software designer often leaves the hint or clue with respect to the reason or rationale for each module or component as the comment on software when he/she designs and implements the software. The comment is utilized for structuring: software into a tree and CBR (Case-Based Reasoning) is carried out based on the tree structure for enabling; the effective re-use of past software with respect to the structure and organization of software in the uppser stream in design. A prototype system has been implemented with the proposed support method and experiments have been carried out to investigate the effectiveness of our approach., 2000年, 311, 315, 研究論文(国際会議プロシーディングス)
査読あり, 日本語, 電子情報通信学会論文誌. A, 基礎・境界, 一般社団法人電子情報通信学会, 権限・義務・責任・知識を考慮した緊急時のコミュニケーションモデル, 小磯 貴史; 吉田 哲也; 才脇 直樹; 西田 正吾, 大規模災害に対処する防災システムを考えるとき, 適切な部署に,適切な情報が適切なタイミングで届くコミュニケーシヨンの実現が重要である. 本論文では, 階層型組織での緊急時の意思決定について分析するとともに, 組織の各構成員のもつ権限・義務・責任・知識に注目することにより, 人的構造を考慮した緊急時のコミュニケーションモデルを提案する. 更にこのコミュニケーションモデルのプラント制御への適用例を示すとともに, このモデルの適用可能性として緊急時のコミユニケーシヨン支援システムへの応用や, 組織形態のコミュニケーシヨンの視点から見た評価についても論じる., 1999年03月, 82, 3, 445, 453, 研究論文(学術雑誌)
査読あり, 英語, NEW DIRECTIONS IN ROUGH SETS, DATA MINING, AND GRANULAR-SOFT COMPUTING, Evolving granules for classification for discovering difference in the usage of words, T Yoshida; T Kondo; S Nishida, This paper proposes an evolutionary approach for discovering difference in the usage of words to facilitate collaboration among people. When people try to communicate their concepts with words, the difference in the meaning and usage of words can lead to misunderstanding in communication, which can hinder their collaboration. In our approach each granule of knowledge in classification from users is structured into a decision tree so that difference in the usage of word can be discovered as the difference in the structure of tree. By treating each granule of classification knowledge as an individual in Genetic Algorithm (GA), evolution is carried out with respect to the classification efficiency of each individual and diversity as a population so that difference in the usage of words will emerge as the difference in the structure of decision tree. Experiments were carried out on motor diagnosis cases with artificially encoded difference in the usage of words and the result shows the effectiveness of our evolutionary approach., 1999年, 1711, 366, 374, 研究論文(学術雑誌)
査読あり, 英語, IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Learning the Balance between Exploration and Exploitation via Reward, 1999年, 82-EA, 11, 2538, 2545, 研究論文(学術雑誌)
査読あり, 日本語, 情報処理学会論文誌, 一般社団法人情報処理学会, 複数の人間における概念相違検出のためのインタフェースの設計(<特集>次世代ヒューマンインタフェース・インタラクション), 近藤 輝幸; 吉田 哲也; 西田 正吾, 複数の人間が参加して共同作業を行う場合には, 異なる背景を持つ人間の間に存在する概念の相違が大きな問題となることが考えられる.本稿では, このような複数の人間における概念の相違について考察するとともに, 同じシンボルを異なる意味に用いたり, 異なるシンボルを同じ意味に用いたりするケースを対象に, ユーザの知識を表す決定木を用いて複数の人間における概念の相違を検出する方法について述べ, プロトタイプを作成し評価する., 1998年05月, 39, 5, 1195, 1202, 研究論文(学術雑誌)
査読あり, 英語, IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES, A cooperation method via metaphor of explanation, T Yoshida; K Hori; S Nakasuka, This paper proposes a new method to improve cooperation in concurrent systems within the framework of Multi-Agent Systems (MAS). Since subsystems work concurrently, achieving appropriate cooperation among them is important to improve the effectiveness of the overall system. When subsystems are modeled as agents, it is easy to explicitly deal with the interactions among them since they can be modeled naturally as communication among agents with intended information. Contrary to previous approaches which provided the syntax of communication protocols without semantics, we focus on the semantics of cooperation in MAS and aim at allowing agents to exploit the communicated information for cooperation. This is attempted by utilizing more coarse-grained communication based on the different perspective for the balance between Formality and richness of communication contents so that each piece of communication contents can convey more meaningful information in application domains. In our approach agents cooperate each other by giving Feedbacks based on the metaphor of explanation which is widely used in human interactions, in contrast to previous approaches which use direct orders given by the leader based on the pre-defined cooperation strategies. Agents show the difference between the proposal and counterproposals for it, which are constructed with respect to the Former and given as the feedbacks in the easily understandable terms For the receiver. From the comparison of proposals agents retrieve the information on which parts are agreed and disagreed by the relevant agents, and reflect the analysis in their following behavior. Furthermore, communication contents are annotated by agents to indicate the degree of importance in decision making for them, which contributes to making explanations or feedbacks more understandable. Our cooperation method was examined through experiments on the design of micro satellites and the result showed that it was effective to some extent to facilitate cooperation among agents., 1998年04月, E81A, 4, 576, 585, 研究論文(学術雑誌)
査読あり, 日本語, 情報処理学会論文誌, 複数の人間における概念相違検出のためのインターフェースの設計, 近藤, 1998年, 39, 5, 1195, 1202, 研究論文(学術雑誌)
査読あり, 英語, INTERNATIONAL CONFERENCE ON MULTI-AGENT SYSTEMS, PROCEEDINGS, A reinforcement learning approach to cooperative problem solving, T Yoshida; K Hori; S Nakasuka, We propose an extension of reinforcement learning methods to cooperative problem solving in multi-agent systems. Exploiting multiple agents for complex problems is promising, however, learning is necessary since complete domain knowledge is rarely available. The temporal difference algorithm is applied in each agent to learn a heuristic evaluation of states. Besides the reward for solutions produced by agents, we define the reward for coherence as a. whole and exploit them to facilitate cooperation among agents for global problem solving. We evaluate the method by experiments on the satellite design problem. The result shows that our method enables agents to learn to cooperate as well as to learn individual heuristics within one framework. Especially, agents themselves learn to take the appropriate balance between exploration and exploitation in problem solving, which is known to greatly affect the performance. It also suggests the possibility of controlling the global behavior of multi-agent systems via rewards in reinforcement learning., 1998年, 479, 480, 研究論文(国際会議プロシーディングス)
査読あり, 英語, ARTIFICIAL INTELLIGENCE IN REAL-TIME CONTROL 1997, Interactive interfaces to detect conceptual difference for group knowledge acquisition, S Nishida; T Yoshida; T Kondo, Conceptual difference is a serious problem in group knowledge acquisition systems, especially when different people with different background participate in a group. This paper deals with conceptual difference and proposes a method to detect it in the cases that different symbols are used as the same meaning and/or same symbols are used as the different meanings. In section 2, conceptual difference is defined and system architecture for detection is described. In section 3, detecting algorithm is designed, and then a prototype system and its evaluation are discussed in section 4. Copyright (C) 1998 IFAC., 1998年, 177, 181, 研究論文(国際会議プロシーディングス)
査読あり, 日本語, 日本建築学会計画系論文集, 日本建築学会, 歴史的風土特別保存地区における民家主屋の外観意匠類型化とその屋敷構えとの関係 - 明日香村の奥山・飛鳥・川原・野口・岡・島庄の六大字を事例として -, 山本 直彦; 平尾 和洋; 吉田哲也; 室崎 千重, Asuka village in Nara prefecture is well-known for its richness in historic remains originated from Asuka period. This paper aims to clarify the exterior design composition of rural main houses in six village sections in Asuka village. Based on the data collected in these village sections, Items/Categories list is compiled. Using this IC list data, as a result of multiple correspondence analysis followed by cluster analysis, samples are classified into 15 clusters. The analysis next searches the relation between house compound types and exterior design patterns of main houses. Finally, the paper tried to figure out each village section’s characteristic., 2022年07月, 87, 797, 1271, 1281, 研究論文(学術雑誌), 10.3130/aija.87.1271
査読あり, その他, Journal of Fiber Bioengineering and Informatics, Generation of Approximate Weave Diagrams via Warp Pick Up Assignment, Tetsuya Yoshida, 2022年07月, 15, 2, 79, 90, 研究論文(学術雑誌)
査読あり, その他, 芸術科学会論文誌, ねじり折りに対するモジュールに基づく花紋スモッキングの組合せ, 吉田哲也,夛田 美沙, 2022年06月, 21, 2, 77, 86, 研究論文(学術雑誌)
査読あり, その他, 情報処理学会論文誌: 数理モデル化と応用, 綜絖枠数の制約下での織物組織図の近似, 吉田哲也, 2022年07月, 15, 3, 90, 96, 研究論文(学術雑誌)
査読あり, その他, 情報処理学会論文誌:数理モデル化と応用, 呼吸の特徴量を用いた心拍間隔の欠損補完, 野村涼子; 吉田哲也, 2022年07月, 15, 3, 11, 18, 研究論文(学術雑誌)
査読あり, その他, Advanced Biomedical Engineering, A Missing RR Interval Complement Method Based on Respiratory Features, Ryoko Nomura; Tetsuya Yoshida, 2022年12月, 11, 237, 248, 研究論文(学術雑誌), 10.14326/abe.11.237
査読あり, 英語, IEEE RO-MAN 2000: 9TH IEEE INTERNATIONAL WORKSHOP ON ROBOT AND HUMAN INTERACTIVE COMMUNICATION, PROCEEDINGS, A method for supporting web page design based on impression of web page, M Watanabe; T Yoshida; N Saiwaki; S Nishida, This paper proposes an approach for web page design support by focusing: on the impression of web page with respect to the selection of color and font. Previous approaches for web page design support often focus on the automation of HTML coding: and the pasting of graphics and icons. However, the selection of color and font also greatly affect the impression of web page even with the same (information) content. The implemented system with the proposed support method tries to acquire the taste or preference of the user from the web pages which are selected by him/her during net surfing; and utilize it for web page design with the prefered impression., 2000年, 13, 17, 研究論文(国際会議プロシーディングス)
査読あり, 英語, SMC '97 CONFERENCE PROCEEDINGS - 1997 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS, VOLS 1-5, Detection algorithm for conceptual difference among different persons - as an example of emergence in the human interaction, T Yoshida; T Kondo; N Saiwaki; S Nishida, We consider emergence in the human interaction by taking the discovery of conceptual difference as an example. Conceptual difference is a serious problem in group knowledge acquisition systems, especially when different people with different background participate in a group. We deal with the conceptual difference, in which different symbols are used to express the same meaning and/or the same symbol is used to express different meanings. We propose a method to detect such types of conceptual difference by utilizing decision trees which are constructed with ID3 algorithm and GA (Genetic Algorithm).
Our method is discussed in terms of the realization of emergence in the human interaction. We tackle the ''vague'' concept of emergence in the human interaction by constructing ''concrete'' structures to express conceptual difference. Our method ensures the variety in the conceptual structures by utilizing GA to explore various possibilities. The various structures constructed by the system play the role of pointing out conceptual difference from another viewpoint, which is different from those of the members in the group. This function enables the members to understand each other as well as their own concepts, which is expected to contribute to hitting new ideas as a group and to promoting the activity of the group., 1997年, 1216, 1221, 研究論文(国際会議プロシーディングス)
査読あり, 英語, DESIGN OF COMPUTING SYSTEMS: SOCIAL AND ERGONOMIC CONSIDERATIONS, Interactive interfaces to detect conceptual difference for knowledge acquisition, T Kondo; N Saiwaki; T Yoshida; S Nishida, Conceptual difference is a serious problem in group knowledge acquisition systems, especially when different people with different background participate in a group. In our former research, each a peace of knowledge held by people is expressed by a single decision tree, and conceptual difference is detected by checking the difference. Sometimes it occurs that the attributes and the values with conceptual difference do not appear in decision trees and fails to detect conceptual difference. In this paper, we propose a method to solve this problem by using decision trees with diverse structures which are produced by our genetic algorithm., 1997年, 21, 303, 306, 研究論文(国際会議プロシーディングス)
その他, Intelligent Data Analysis, Adaptive Ripple Down Rules method based on minimum description length principle., Tetsuya Yoshida; Takuya Wada; Hiroshi Motoda; Takashi Washio, 2004年, 8, 3, 239, 265, 研究論文(学術雑誌)
その他, Advances in Knowledge Discovery and Data Mining(PAKDD), Classifier Construction by Graph-Based Induction for Graph-Structured Data., Warodom Geamsakul; Takashi Matsuda; Tetsuya Yoshida; Hiroshi Motoda; Takashi Washio, 2003年, 52, 62, 研究論文(国際会議プロシーディングス), 10.1007/3-540-36175-8_6
その他, Discovery Science, Performance Evaluation of Decision Tree Graph-Based Induction., Warodom Geamsakul; Takashi Matsuda; Tetsuya Yoshida; Hiroshi Motoda; Takashi Washio, 2003年, 128, 140, 研究論文(国際会議プロシーディングス), 10.1007/978-3-540-39644-4_12
その他, PRICAI 2002: Trends in Artificial Intelligence(PRICAI), Knowledge Discovery from Structured Data by Beam-Wise Graph-Based Induction., Takashi Matsuda; Hiroshi Motoda; Tetsuya Yoshida; Takashi Washio, 2002年, 255, 264, 研究論文(国際会議プロシーディングス), 10.1007/3-540-45683-X_29
その他, PRICAI 2002: Trends in Artificial Intelligence(PRICAI), Case Generation Method for Constructing an RDR Knowledge Base., Keisei Fujiwara; Tetsuya Yoshida; Hiroshi Motoda; Takashi Washio, 2002年, 228, 237, 研究論文(国際会議プロシーディングス), 10.1007/3-540-45683-X_26
その他, PRICAI 2002: Trends in Artificial Intelligence(PRICAI), Extension of the RDR Method That Can Adapt to Environmental Changes and Acquire Knowledge from Both Experts and Data., Takuya Wada; Tetsuya Yoshida; Hiroshi Motoda; Takashi Washio, 2002年, 218, 227, 研究論文(国際会議プロシーディングス), 10.1007/3-540-45683-X_25
その他, Discovery Science, Mining Patterns from Structured Data by Beam-Wise Graph-Based Induction., Takashi Matsuda; Hiroshi Motoda; Tetsuya Yoshida; Takashi Washio, 2002年, 422, 429, 研究論文(国際会議プロシーディングス), 10.1007/3-540-36182-0_44
査読あり, その他, Journal of Fiber Bioengineering and Informatics, Petaloid Folding of General Triangles for Combinatorial Smocking Design, Yoshida, T, 2022年09月, 15, 3, 145, 156
査読あり, 英語, Textile Research Journal, A method for approximating weave diagrams under heald frame constraint, Tetsuya Yoshida, Woven fabric is produced by interlacing the warp yarn and the weft yarn, and their relationship on woven fabric is represented as a weave diagram. In order to produce woven fabric with looms, the necessary number of heald frames can be calculated based on the weave diagram for the woven fabric. However, when the number of available heald frames is limited, it is difficult to produce complex woven fabric. To alleviate this problem, this article proposes a method for approximating weave diagrams under heald frame constraint in terms of the alternating optimization in machine learning. By defining the objective function based on the matrix representation of weave diagram, an alternating optimization method is proposed to generate an approximate weave diagram under the specified number of heald frames. The proposed method is implemented with R language, and evaluated in terms of the performance of the proposed algorithm and woven fabric. Experimental results indicate that the proposed method is effective in generating approximate weave diagrams so that they can be fabricated using looms with a limited number of heald frames., 2023年07月20日, 93, 15-16, 3527, 3537, 研究論文(学術雑誌), 10.1177/00405175221149684
査読あり, その他, 日本建築学会計画系論文集, 修景に向けた民家の外観意匠類型化手法と景観評価, 鈴木 裕子; 山本 直彦; 吉田哲也, 2023年10月, 88, 812, 2751, 2761
査読あり, その他, Japan Architectural Review, Typology on exterior design of townhouses in a World Cultural Heritage Site of a Nepali City, Asuka Hamaoka; Naohiko Yamamoto; Tetsuya Yoshida; Anri Miyauchi; Masaya Masui; Yoichi Mukai, 2023年10月, 6
査読あり, その他, 芸術科学会論文誌, 模様の回転と帯の長さに基づく花紋スモッキングの動的可視化, 吉田 哲也; 中西彩子, 2023年11月, 22, 4, 10:1, 10:11
査読無し, 日本語, 人工知能学会誌, 社団法人人工知能学会, 構造データからのアクティブマイニング(<特集>アクティブマイニング), 元田 浩; ホー ツー バオ; 鷲尾 隆; 矢田 勝俊; 吉田 哲也; 大原 剛三, 2005年03月01日, 20, 2, 172, 179
査読無し, 英語, 情報処理学会研究報告. ICS, [知能と複雑系], 一般社団法人情報処理学会, Decision Tree Graph-Based Inductionによる肝炎データからの診断知識の抽出(9月14日)(<特集>「アクティブマイニング」及び一般), ジアムサクン ワロドム; 吉田 哲也; 大原 剛三; 元田 浩; 鷲尾 隆, Decision Tree Graph-Based Induction (DT-GBI法)はグラフ構造データから決定木を構築する手法である。 DT-GBI法では、部分グラフ(分類に効果的なパターン)を逐次ペア拡張(チャンキング)により抽出し、決定木の各ノードでの分岐パターンとして使用する。DT-GBI法を千葉大学付属病院より提供された肝炎データセットに適用した。実験1では、線維化の段階(程度)をクラスとし、血液検査の時系列のみで第4段階(肝硬変)の患者とそれ以外の段階の患者を分類する決定木を構築し、第2実験では、肝炎の型(BまたはC)をクラスとし、第1実験と同様に肝炎の型を分類する決定木を構築する。初期実験の結果で得られた決定木,分類に効く検査パターンおよび予測精度を報告する。, 2003年09月14日, 2003, 90, 53, 58
査読無し, 英語, 知識ベ-スシステム研究会, 人工知能学会, Extracting Diagnostic Knowledge from Hepatitis Data by Decision Tree Graph-Based Induction (小特集 「アクティブマイニング」および一般), Geamsakul Warodom; 吉田 哲也; 大原 剛三, 2003年09月14日, 61, 0, 53, 58
査読無し, 英語, 電子情報通信学会技術研究報告. AI, 人工知能と知識処理, 一般社団法人電子情報通信学会, Decision Tree Graph-Based Inductionによる肝炎データからの診断知識の抽出(<特集>「アクティブマイニング」及び一般), ジアムサクン ワロドム; 吉田 哲也; 大原 剛三; 元田 浩; 鷲尾 隆, Decision Tree Graph-Based Induction (DT-GBI法)はグラフ構造データから決定木を構築する手法である。DT-GBI法では、部分グラフ(分類に効果的なパターン)を逐次ペア拡張(チャンキング)により抽出し、決定木の各ノードでの分岐パターンとして使用する。DT-GBI法を千葉大学付属病院より提供された肝炎データセットに適用した。実験1では、線維化の段階(程度)をクラスとし、血液検査の時系列のみで第4段階(肝硬変)の患者とそれ以外の段階の患者を分類する決定木を構築し、第2実験では、肝炎の型(BまたはC)をクラスとし、第1実験と同様に肝炎の型を分類する決定木を構築する。初期実験の結果で得られた決定木,分類に効く検査パターンおよび予測精度を報告する。, 2003年09月07日, 103, 304, 53, 58
査読無し, 日本語, 人工知能学会誌, 社団法人人工知能学会, MLnet(Web探訪), 吉田 哲也, 2003年07月01日, 18, 4, 470, 470
査読無し, 日本語, 情報処理学会研究報告. ICS, [知能と複雑系], 一般社団法人情報処理学会, Decision Tree - Graph-Based Inductionの機能拡張, ジアムサクン ワロドム; 松田 喬; 吉田 哲也; 元田 浩; 鷲尾 隆, Graph-Based Induction (GBI法)は,逐次的ペア拡張によってグラフ構造データから類型パターンを抽出する機械学習手法の一つである.一方,決定木はデータ分類に有効な手段であり,理解しやすいルールが得られるという利点があるが,決定木を構築するにはデータを属性-属性値として表現する必要がある.本稿は,GBI法を用いて,グラフ構造データに対して分類器(決定木)を構築する手法,Decision Tree - Graph-Based Induction (DT-GBI法)を提案する.この手法は,オンラインで属性(分類に有効な部分グラフ)をGBI法により生成しながら決定木を構築するという特徴を持つ.DT-GBI法の性能をUCI RepositoryからのDNAデータセットに対する実験で評価し,DT-GBI法がグラフ構造データに対して分類器を構築する効率的な手法であることを示す., 2003年03月13日, 2003, 30, 93, 98
査読無し, 日本語, 情報処理学会研究報告. ICS, [知能と複雑系], 一般社団法人情報処理学会, AGMによる3次元構造と生理活性の相関解析, 西村 芳男; 鷲尾 隆; 吉田 哲也; 元田 浩; 猪口 明博; 岡田 孝, AGMアルゴリズムは,グラフ構造データベース中のグラフデータに共通して多く現れる部分構造パターンを完全探索で抽出するアルゴリズムである.AGMアルゴリズムは頂点と辺にラベルを持つ一般のグラフデータを扱うことができ,グラフのトポロジーを解析できる.本稿では,頂点が3次元座標で表されたグラフデータに対し,各頂点間の距離を計算して辺ラベルに情報を追加する前処理を行い,AGMアルゴリズムで3次元構造の解析を行う手法を提案する.また,実データによって3次元構造解析を行い,生理活性の相関を調査する., 2003年03月13日, 2003, 30, 99, 103
査読無し, 日本語, 知識ベ-スシステム研究会, 人工知能学会, Decision Tree-Graph-Based Inductionの機能拡張 (知識ベースシステム研究会(第60回) 人工知能基礎論研究会(第52回) 小特集:「データマイニング」および一般) -- (文部科学省科学研究費特定領域研究 情報洪水時代におけるアクティブマイニングの実現), Geamsakul Warodom; 松田 喬; 吉田 哲也, 2003年03月13日, 60, 0, 93, 98
査読無し, 日本語, 知識ベ-スシステム研究会, 人工知能学会, AGMによる3次元構造と生理活性の相関解析 (知識ベースシステム研究会(第60回) 人工知能基礎論研究会(第52回) 小特集:「データマイニング」および一般) -- (文部科学省科学研究費特定領域研究 情報洪水時代におけるアクティブマイニングの実現), 西村 芳男; 鷲尾 隆; 吉田 哲也, 2003年03月13日, 60, 0, 99, 103
査読無し, 日本語, 電子情報通信学会技術研究報告. AI, 人工知能と知識処理, 一般社団法人電子情報通信学会, Decision Tree-Graph-Based Inductionの機能拡張(<特集>「アクティブマイニング」及び一般 : 文部科学省科学研究費特定領域研究「情報洪水時代におけるアクティブマイニングの実現」公開シンポジウム), ジアムサクン ワロドム; 松田 喬; 吉田 哲也; 元田 浩; 鷲尾 隆, Graph-Based Induction(GBI法)は,逐次的ペア拡張によってグラフ構造データから類型パターンを抽出する機械学習手法の一つである.一方,決定木はデータ分類に有効な手段であり,理解しやすいルールが得られるという利点があるが,決定木を構築するにはデータを属性-属性値として表現する必要がある.本稿は,GBI法を用いて,グラフ構造データに対して分類器(決定木)を構築する手法,Decision Tree-Graph-Based Induction(DT-GBI法)を提案する.この手法は,オンラインで属性(分類に有効な部分グラフ)をGBI法により生成しながら決定木を構築するという特徴を持つ.DT-GBI法の性能をUCI RepositoryからのDNAデータセットに対する実験で評価し,DT-GBI法がグラフ構造データに対して分類器を構築する効率的な手法であることを示す., 2003年03月07日, 102, 710, 41, 46
査読無し, 日本語, 電子情報通信学会技術研究報告. AI, 人工知能と知識処理, 一般社団法人電子情報通信学会, AGMによる3次元構造と生理活性の相関解析(<特集>「アクティブマイニング」及び一般 : 文部科学省科学研究費特定領域研究「情報洪水時代におけるアクティブマイニングの実現」公開シンポジウム), 西村 芳男; 鷲尾 隆; 吉田 哲也; 元田 浩; 猪口 明博; 岡田 孝, AGMアルゴリズムは,グラフ構造データベース中のグラフデータに共通して多く現れる部分構造パターンを完全探索で抽出するアルゴリズムである.AGMアルゴリズムは頂点と辺にラベルを持つ一般のグラフデータを扱うことができ,グラフのトポロジーを解析できる.本稿では,頂点が3次元座標で表されたグラフデータに対し,各頂点間の距離を計算して辺ラベルに情報を追加する前処理を行い,AGMアルゴリズムで3次元構造の解析を行う手法を提案する.また,実データによって3次元構造解析を行い,生理活性の相関を調査する., 2003年03月07日, 102, 710, 47, 51
査読無し, 日本語, マテリアルインテグレ-ション, ティ-・アイ・シ-, グラフ構造データからのマイニング (特集 大阪大学産業科学研究所 マテリアルインテグレーション--材料・生体・情報の融合を目指して(2)), 元田 浩; 鷲尾 隆; 吉田 哲也, 2002年08月, 15, 8, 59, 62
査読無し, 日本語, マテリアルインテグレ-ション, ティ-・アイ・シ-, 専門家とデータの両方からの統一的知識獲得 (特集 大阪大学産業科学研究所 マテリアルインテグレーション--材料・生体・情報の融合を目指して(2)), 吉田 哲也; 元田 浩; 鷲尾 隆, 2002年08月, 15, 8, 63, 66
査読無し, 日本語, 情報処理学会研究報告. ICS, [知能と複雑系], 一般社団法人情報処理学会, Apriori-based Graph Miningアルゴリズムの高速化, 西村 芳男; 鷲尾 隆; 吉田 哲也; 元田 浩; 猪口 明博, Apriri-based Graph Mining(AGM)アルゴリズムはAprioriアルゴリズムをグラフ構造データに拡張したアルゴリズムであり,グラフ構造データから特徴的なパターンを抽出する.提案手法は,AGMアルゴリズムの多頻度グラフの候補生成において異なった条件を使用し,その候補生成の計算時間を短縮する.提案手法を実装したシステムを構築し,人工データによる性能評価と提案手法の特徴分析を行う.さらに,提案手法を実際のデータに適用し,AGMアルゴリズム全体の計算時間も高速化できることを示す., 2002年05月23日, 2002, 45, 11, 16
査読無し, 日本語, 情報処理学会研究報告. ICS, [知能と複雑系], 一般社団法人情報処理学会, グラフ構造に着目した肝炎データからの知識発見, 松田 喬; 吉田 哲也; 元田 浩; 鷲尾 隆, 欠損値の非常に多いデータは表形式で取り扱うのに不向きであるため,欠損値にロバストなグラフ構造を用いて肝炎データの解析を試みた結果を報告する., 2002年05月23日, 2002, 45, 67, 72
査読無し, 日本語, 知識ベ-スシステム研究会, 人工知能学会, Apriori-based Graph Mining アルゴリズムの高速化 (テーマ:「アクティブマイニング」および一般), 西村 芳男; 鷲尾 隆; 吉田 哲也, 2002年05月23日, 56, 0, 11, 16
査読無し, 日本語, 知識ベ-スシステム研究会, 人工知能学会, グラフ構造に着目した肝炎データからの知識発見 (テーマ:「アクティブマイニング」および一般), 松田 喬; 吉田 哲也; 元田 浩, 2002年05月23日, 56, 0, 67, 72
査読無し, 日本語, ヒューマンインタフェース学会論文誌, ヒュ-マンインタフェ-ス学会, ユーザ参加型設計のための事例を利用した意図理解支援インタフェース, 山岡 孝行; 辻野 克彦; 吉田 哲也; 西田 正吾, 2000年05月01日, 2, 2, 87, 95
査読無し, その他, Adaptive Hypermedia System for Supporting Information Providers to Direct Users through Hyperspace, 2000年, 120-C, 11, 1720, 1731
査読無し, その他, Supporting Mutual Understanding in Participatory Design Using Cases, 2000年, 2, 2, 87, 95
査読無し, 英語, IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES, IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, Learning the balance between exploration and exploitation via reward, T Yoshida; K Hori; S Nakasuka, This paper proposes a new method to improve cooperation in concurrent systems within the framework of Multi-Agent Systems (MAS) by utilizing reinforcement learning. When subsystems work independently and concurrently, achieving appropriate cooperation among them is important to improve the effectiveness of the overall system. Treating subsystems as agents makes it easy to explicitly deal with the interactions among them since they can be modeled naturally as communication among agents with intended information. In our approach agents try to learn the appropriate balance between exploration and exploitation via reward, which is important in distributed and concurrent problem solving in general. By focusing on how to give reward in reinforcement learning, not the learning equation, two kinds of reward are defined in the context of cooperation between agents, in contrast to reinforcement learning within the framework of single agent. In our approach reward for insistence by individual agent contributes to facilitating exploration and reward for concessi:on to other agents contributes to facilitating exploitation. Our cooperation method was examined through experiments on the design of micro satellites and the result showed that it was effective to some extent to facilitate cooperation among agents by letting agents themselves learn the appropriate balance between insistence and concession. The result also suggested the possibility of utilizing the relative magnitude of these rewards as a new control parameter in MAS to control the overall behavior of MAS., 1999年11月, E82A, 11, 2538, 2545
査読無し, 日本語, 知識ベ-スシステム研究会, 人工知能学会, 決定木間のリンクを利用した概念相違発見手法 (テーマ:「インターネットとAI」および一般), 大西 健介; 吉田 哲也; 西田 正吾, 1999年03月, 43, 69, 74
査読無し, 日本語, 知識ベ-スシステム研究会, 人工知能学会, 検索キーワードの補完情報を利用した情報獲得支援 (テーマ:「インターネットとAI」および一般), 新開 大樹; 吉田 哲也; 西田 正吾, 1999年03月, 43, 105, 110
査読無し, その他, Proc. of the Third Pacific-Asia Conference on Knowledge Discovery and Data Mining, Beijing in China, 1999, Discovering Conceptual Differences among Different People via Diverse Structures, YOSHIDA T., 1999年, 494, 498
査読無し, 日本語, 電子情報通信学会論文誌. A, 基礎・境界, 一般社団法人電子情報通信学会, 権限・義務・責任・知識を考慮した緊急時のコミュニケーションモデル, 小磯 貴史; 吉田 哲也; 才脇 直樹; 西田 正吾, 大規模災害に対処する防災システムを考えるとき, 適切な部署に,適切な情報が適切なタイミングで届くコミュニケーシヨンの実現が重要である. 本論文では, 階層型組織での緊急時の意思決定について分析するとともに, 組織の各構成員のもつ権限・義務・責任・知識に注目することにより, 人的構造を考慮した緊急時のコミュニケーションモデルを提案する. 更にこのコミュニケーションモデルのプラント制御への適用例を示すとともに, このモデルの適用可能性として緊急時のコミユニケーシヨン支援システムへの応用や, 組織形態のコミュニケーシヨンの視点から見た評価についても論じる., 1999年, 82A, 3, 445, 453
査読無し, 日本語, 全国大会講演論文集, 一般社団法人情報処理学会, メタデータとユーザモデルを利用したハイパーメディアにおける動的リンク方式, 土方 嘉徳; 吉田 哲也; 西田 正吾, 1998年10月05日, 57, 3, 105, 106
査読無し, 英語, IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES, IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, A cooperation method via metaphor of explanation, T Yoshida; K Hori; S Nakasuka, This paper proposes a new method to improve cooperation in concurrent systems within the framework of Multi-Agent Systems (MAS). Since subsystems work concurrently, achieving appropriate cooperation among them is important to improve the effectiveness of the overall system. When subsystems are modeled as agents, it is easy to explicitly deal with the interactions among them since they can be modeled naturally as communication among agents with intended information. Contrary to previous approaches which provided the syntax of communication protocols without semantics, we focus on the semantics of cooperation in MAS and aim at allowing agents to exploit the communicated information for cooperation. This is attempted by utilizing more coarse-grained communication based on the different perspective for the balance between Formality and richness of communication contents so that each piece of communication contents can convey more meaningful information in application domains. In our approach agents cooperate each other by giving Feedbacks based on the metaphor of explanation which is widely used in human interactions, in contrast to previous approaches which use direct orders given by the leader based on the pre-defined cooperation strategies. Agents show the difference between the proposal and counterproposals for it, which are constructed with respect to the Former and given as the feedbacks in the easily understandable terms For the receiver. From the comparison of proposals agents retrieve the information on which parts are agreed and disagreed by the relevant agents, and reflect the analysis in their following behavior. Furthermore, communication contents are annotated by agents to indicate the degree of importance in decision making for them, which contributes to making explanations or feedbacks more understandable. Our cooperation method was examined through experiments on the design of micro satellites and the result showed that it was effective to some extent to facilitate cooperation among agents., 1998年04月, E81A, 4, 576, 585
査読無し, その他, Software Design Support by Utilizing Comment, 1998年, 113, 118
査読無し, その他, 7th IFAC SYMPOSIUM on Analysis, Design and Evaluation of Man-Machine Systems, Facilitate Cooperation of Humans and Machines from the viewpoint of Multi-Agent Systems, 1998年, 485, 490
査読無し, 英語, DISCOVERY SCIENCE, SPRINGER-VERLAG BERLIN, Discovering conceptual differences among people from cases, T Yoshida; T Kondo, We propose a method for discovering conceptual differences (CD) among people from cases. In general different people seem to have different ways of conception and thus can have different concepts even on the same thing.
Removing CD seems especially important when people with different backgrounds and knowledge carry out collaborative works as a group; otherwise they cannot communicate ideas and establish mutual understanding even on the same thing. In our approach knowledge of users is structured into decision trees so that differences in concepts can be discovered as the differences in the structure of trees. Based on the candidates suggested by the system with our discovering algorithms, the users then discuss each other on differences in their concepts and modify them to reduce the differences. CD is gradually removed by repeating the interaction between the system and users. Experiments were carried out on the cases for motor diagnosis with artificially encoded CD. Admittedly our approach is simple, however, the result shows that our approach is effective to some extent as the first step toward dealing with the issue of CD among people., 1998年, 1532, 162, 173
査読あり, その他, 7th IEEE International Workshop on Robot and Human Communication(ROMAN'98), An Extension of Reinforcement Learning in Multi-Agent Systems, 1998年, 221, 226, 研究発表ペーパー・要旨(国際会議)
査読無し, 英語, INTELLIGENT AUTONOMOUS SYSTEMS, I O S PRESS, A cooperation method to exploit design rationales for agents, T Yoshida; K Hori; S Nakasuka, This paper proposes a new method to improve cooperation in Multi-Agent Systems (MAS) from the viewpoint of Design Rationale (DR). In our approach agents cooperate each other by giving feedbacks on the proposal from each agent indirectly, in contrast to previous approaches which use direct orders given by the leader based on the pre-defined cooperation strategies. Indirect cooperation is realized by constructing and giving counter-proposals from relevant agents on the proposal. Feedbacks given as the counter-proposals play the role of pointing out insufficient portions in the original one from different perspectives toward the agreement among agents. Based on the comparison of proposals agents retrieve the information on which parts are agreed and disagreed by the relevant agents, and reflect the analysis in their following behavior. Furthermore, communication contents are annotated by agents to indicate the degree of importance in decision making for them, which contributes to making feedbacks more understandable. Our cooperation method was examined through experiments on the design of micro satellites and the result showed that it was effective to some extent to facilitate cooperation among agents., 1998年, 647, 654
査読無し, 英語, 3RD ASIA PACIFIC COMPUTER HUMAN INTERACTION, PROCEEDINGS, IEEE COMPUTER SOC, Supporting mutual understanding in collaborative design project, T Yamaoka; K Tsujino; T Yoshida; S Nishida, In collaborative design, participants usually have different backgrounds and standpoints. Due to the differences, it is hard for the participants mutually to understand the intention and they have to make much effort to reach a mutual agreement on the design, then the efficiency of design processes is decreased. In this paper, we propose a framework to support the participants for making the mutual understanding by grasping and showing differences among their intention in collaboration. A system based on the proposed framework has the characteristics: facilitating mutual understanding among participants by presuming their intention and by communicating them with the understandable form, visualizing the difference in concepts among participants, and making progress collaboration in an interactive way., 1998年, 132, 137
査読無し, その他, Design of the Interfaces to Detect Conceptual Differnce among Defferent People, 1998年, 39, 5, 1195, 1202
査読無し, 日本語, 電子情報通信学会技術研究報告. HCS, ヒューマンコミュニケーション基礎, 一般社団法人電子情報通信学会, コミュニケーションの視点からみた緊急時の組織形態の評価, 小磯 貴史; 才脇 直樹; 吉田 哲也; 西田 正吾, 本稿では, 筆者らが以前に提案したコミュニケーションモデルの適用可能性を考えるにあたって, 管理組織の形態を考慮する際の, 緊急時におけるコミュニケーションの視点から見た評価について考えた. 組織形態の評価は, 主に社会科学の分野で定性的な議論がなされているが, 本研究ではそれらの評価を定量的に扱えるツールとして利用することを目指す. まずは緊急時のコミュニケーションモデルについて簡単に述べる. そして, 組織形態の評価手法をいくつか紹介し, そのプロトタイプについて説明する., 1997年06月13日, 97, 99, 5, 12
査読無し, その他, 11th European Conference on Artificial Intelligence, Coloured rippling : an extension of a theorem proving heuristics, 1994年, 85, 89
日本語, 研究報告数理モデル化と問題解決(MPS), ダイバージェンスに基づくNMFを用いた転移学習, 木村 圭吾; 吉田 哲也, 本稿では,トピックグラフに基づく転移学習法を拡張し,一般化 KL (Kullback-Leibler) ダイバージェンスに基づく NMF (Non-negative Matrix Factorization) を用いた転移学習法を提案する.ダイバージェンスを通じた転移学習の確率的な解釈を目指して,フロベニウスノルムに基いてトピックの関係 (トピックグラフ) を活用する転移学習法を拡張し,転移学習を一般化KLダイバージェンスに基づく最適化問題として定式化する.最適化規準に対する補助関数を定義し,補助関数から最適化アルゴリズムを導出し,その収束性を示す.提案法を文書クラスタリングに適用し,他手法との比較を通じて提案法の有効性を示す.特に,提案法による転移学習を通じてダイバージェンスを用いた場合でも精度向上を実現できることを示す.We propose a topic graph based transfer learning method based on Non-negative Matrix Factorization (NMF) with generalized Kullback-Leibler (KL) divergence. In this paper we extend the previous NMF based transfer learning method by utilizing generalized KL divergence based NMF so that better probabilistic interpretation can be obtained with the divergence. The proposed method is formalized as the minimization of an objective function under the divergence, and an auxiliary function for the objective function is defined. From the auxiliary function, we derive a learning algorithm with multiplicative update rules, which are guaranteed to converge. The proposed method is evaluated in terms of document clustering over several well-known benchmark datasets. Especially, one drawback of generalized KL divergence based NMF algorithms is performance degradation compared with Frobenius based ones. The experimental results show that, by utilizing the topic graph, the proposed method enables to boost up the performance even with KL divergence based NMF through transfer learning., 2011年09月08日, 2011, 3, 1, 6
日本語, 研究報告バイオ情報学(BIO), 相互情報量に基づくクラスタリングに対するグラフモデルとその評価, 吉田 哲也, 本稿では,相互情報量に基づくクラスタリング問題に対するグラフモデルを提案する.相互情報量から導出される定常分布に着想を得たデータ間の類似度関数を定義してデータ集合を辺重み付きグラフとして表現することにより,データが一様分布する場合にはハードクラスタリング問題が提案するグラフモデルにおける組合せ最適化問題に近似できることを示す.提案するグラフモデルを文書クラスタリングでのベンチマークデータである 20 Newsgroup のデータに対して評価し,他手法との比較を通じて提案手法の妥当性と有効性を確認した.We propose a graph model for data clustering based on mutual information. Based on the stationary distribution induced from the problem setting, we propose a similarity function among data objects, and represent the entire objects as an edge-weighted graph. We show that, in hard assignment, the problem can be approximated as a combinatorial problem over the proposed graph when data is uniformly distributed. The proposed approach is evaluated on the text clustering problem over the 20 Newsgroup benchmark data. The results are encouraging and indicate the effectiveness of our approach., 2009年12月10日, 2009, 30, 1, 7
日本語, 人工知能学会全国大会論文集, 人工知能学会, Graph-Based Inductionによる分類学習のための構造データからの属性構築, 松田 喬; 元田 浩; 吉田 哲也, 2002年, 16, 1, 4
日本語, 人工知能学会全国大会論文集, 人工知能学会, 近傍事例のクラス情報のみを利用したRDR知識ベース構築のための事例生成手法と評価, 藤原 啓成; 吉田 哲也; 元田 浩, 2002年, 16, 1, 4
日本語, 人工知能学会全国大会論文集, 人工知能学会, 柔軟な知識ベースシステム構築を目指すDescription Length-based RDR法の人工データによるシミュレーション, 和田 卓也; 吉田 哲也; 元田 浩, 2002年, 16, 1, 4
日本語, 研究報告数理モデル化と問題解決(MPS), 特徴表現のスパース性を考慮したNMF, 木村 圭吾; 吉田 哲也, 本稿では,特徴表現のスパース制約を考慮した NMF (Non-negative Matrix Factorization) を提案する.近年,要素が非負である実行列を,同じく要素が非負である実行列の積として表現する非負値行列分解 (NMF) が注目を集めている.従来の研究では NMF における非負性制約が非零の要素が少ないスパースな特徴表現の学習に寄与すると考えられ,またスパース制約を導入した手法も提案されているが,これまで特徴表現のスパース性は明示的には考慮されてこなかった.本稿では NMF における特徴表現に着目し,特徴表現のスパース性を独立性と相関から定式化し,定式化したスパース性を正則化項として活用する手法を提案する.提案法を文書クラスタリングに適用し,従来法との比較を通じて提案法の有効性を示す.We propose an approach for Non-negative Matrix Factorization (NMF) with sparseness constraints on features. It has been believed that the non-negativity constraint in NMF contributes to making the learned features sparse. In addition, several approaches incorporated additional sparseness constraints, by hoping that the constraints make the features more sparse. However, previous approaches have mostly focused on coefficients, and have not considered the sparsity of features explicitly. Our approach explicitly incorporates the sparsity of features, in terms of independence of features and correlation of features. The proposed notion of sparsity is formalized as regularization terms in the framework of NMF, and learning algorithms with multiplicative update rules are proposed. The proposed approach is evaluated in terms of document clustering over well-known benchmark datasets. Several experiments have been conducted on the datasets, and comparison with other state-of-the-art NMF algorithms is reported. The results are encouraging and show that the proposed approach improves the clustering performance, while sustaining relatively good quality of data approximation., 2011年09月08日, 2011, 2, 1, 6
日本語, 情報科学技術レターズ, FIT(電子情報通信学会・情報処理学会)運営委員会, LD-006 局所更新に基づく逐次的な近傍グラフ構築手法(データベース), Hacid Hakim; 吉田 哲也, 2007年08月22日, 6, 103, 106
日本語, 人工知能学会全国大会論文集, 人工知能学会, 部分グラフを制約とするグラフ構造データからの知識獲得, 茂木 明; 吉田 哲也; ジアムサクン ワロドム, 2004年, 18, 1, 4
日本語, 人工知能学会全国大会論文集, 人工知能学会, Decision Tree-Graph-Based Inductionの探索能力改善, ジアムサクン ワロドム; 松田 喬; 吉田 哲也, 2003年, 17, 1, 4
日本語, 知識ベ-スシステム研究会, 人工知能学会, Decision Tree Graph-based Induction法による肝炎データからの診断知識発見 (特集 「医療及び化学情報マイニング」および一般), Geamsakul Warodom; 吉田 哲也; 大原 剛三, 2004年03月01日, 64, 47, 54
日本語, ヒュ-マンインタフェ-スデザイン研究会, 人工知能学会, 視点の異なる参加者による協同設計の支援手法の提案, 山岡 孝行; 辻野 克彦; 吉田 哲也, 1997年06月, 31, 19, 24