
兼岩 憲
情報・ネットワーク工学専攻 | 教授 |
Ⅰ類(情報系) | 教授 |
- プロフィール:
電気通信大学にて以下の研究課題に従事.
・オントロジーや論理表現の推論に対する機械学習
・知識グラフやグラフデータに対する機械学習
・セマンティックWebのための大規模RDFデータに対する圧縮と高速検索
・ラフ集合データマイニング
・システム設計のための構造的知識(UMLやオントロジーなど)や規則的知識(ルールやプロトコルなど)の矛盾検証
研究者情報
研究キーワード
研究活動情報
受賞
論文
- Predicting from a Different Perspective: A Re-ranking Model for Inductive Knowledge Graph Completion
Yuki Iwamoto; Ken Kaneiwa
ラスト(シニア)オーサー, Proceedings of the 21st Pacific Rim International Conference on Artificial Intelligence (PRICAI 2024), 掲載ページ 299-304, 出版日 2024年11月23日, 査読付
研究論文(国際会議プロシーディングス), 英語 - Efficient Index Structures for Query Solutions on RDF Graphs
Koji Fujiwara; Ken Kaneiwa
ラスト(シニア)オーサー, Proceedings of the 26th IEEE/ACIS International Conference on Computer and Information Science (ICIS 2024-Summer III), 出版日 2024年07月16日, 査読付
英語 - The Completeness of Reasoning Algorithms for Clause Sets in Description Logic ALC
Ken Kaneiwa; Daiki Takahashi
筆頭著者, Knowledge-Based Systems, 286巻, 掲載ページ Article 111405, 出版日 2024年01月28日, 査読付
研究論文(学術雑誌), 英語 - Embedding Models with Inverted-index and Co-occurrence Matrices for Ontology Subsumption Prediction
Yukihiro Shiraishi; Ken Kaneiwa
Proceedings of the 12th International Joint Conference on Knowledge Graphs (IJCKG 2023), 出版日 2023年12月08日, 査読付
研究論文(国際会議プロシーディングス), 英語 - Feature Selection Based on the Complexity of Structural Patterns in RDF Graphs
Ken Kaneiwa; Yota Minami
筆頭著者, International Journal of Data Science and Analytics, 16巻, 出版日 2023年11月10日, 査読付
研究論文(学術雑誌), 英語 - Block-segmentation vectors for arousal prediction using semi-supervised learning
Yuki Odaka; Ken Kaneiwa
ラスト(シニア)オーサー, Applied Soft Computing, Elsevier BV, 142巻, 掲載ページ 110327-110327, 出版日 2023年07月, 査読付
研究論文(学術雑誌), 英語 - 複数の大規模RDFデータセットからの同値関係による集結パス検索
兼岩 憲,山中 佑紀
人工知能学会論文誌, 3838巻, 2号, 掲載ページ D-M53_1-9, 出版日 2023年03月01日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - 重複排除を含むSPARQLクエリの等価変換と計算量
兼岩 憲, 平山 健太
日本データベース学会和文論文誌, 23-J巻, Article No.1号, 出版日 2023年03月, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - 極小RDF推論に基づく記述論理SROIQの概念生成
兼岩 憲,長井拓馬
人工知能学会論文誌, 35巻, 1号, 掲載ページ B-J62_1-13, 出版日 2020年01月01日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - RDFグラフの多様性に対する汎用カーネル関数
荒井大地, 兼岩 憲
ラスト(シニア)オーサー, 人工知能学会論文誌, 33巻, 5号, 掲載ページ B-I12_1-14, 出版日 2018年09月03日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - RDF データに対するグラフパターンマイニング
廣橋美紀, 兼岩 憲
日本データベース学会和文論文誌, 17-J巻, Article No.1号, 出版日 2018年06月28日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - 大規模RDFグラフに対するインデックス付きデータ圧縮と高速検索
兼岩 憲, 藤原浩司
人工知能学会論文誌, 33巻, 2号, 掲載ページ E-H43_1-10, 出版日 2018年03月01日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - RDFグラフの冗長な特徴表現に対するカーネル関数とその高速計算
荒井大地, 兼岩 憲
ラスト(シニア)オーサー, 人工知能学会論文誌, 32巻, 1号, 掲載ページ B-G34_1-12, 出版日 2017年01月13日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - A Logical and Ontological Framework for Compositional Concepts of Objects and Properties
Ken Kaneiwa; Riichiro Mizoguchi; Philip H. P. Nguyen
NEW GENERATION COMPUTING, SPRINGER, 33巻, 2号, 掲載ページ 149-172, 出版日 2015年04月, 査読付, In order to formalize complex and compositional concepts, we propose a logical framework based on an upper ontology constructed from the composition of basic concepts such as properties and objects. In particular, ontologically distinct compositions (called ontological compositions) that are not easily defined by using ISA and PART-OF relations are classified into characterizing, temporal, and spatial compositions (e.g., 'red face' and 'today'). In this paper, we precisely model such ontological compositions by using monadic second-order logic; properties and objects are expressed as predicates, and attributes are expressed as predicates of predicates. The proposed approach provides a novel technique for the classification of attributes as higher-order concepts, and it clarifies illegal compositions of properties and objects and the uniqueness of temporal attributes. Moreover, our composition ontology is described by a set of RDF triples using the metamodeling of concepts in RDF Schema.
研究論文(学術雑誌), 英語 - 大規模RDFグラフのための効率的なクエリ解決
藤原浩司, 兼岩 憲
ラスト(シニア)オーサー, 人工知能学会論文誌, 29巻, 4号, 掲載ページ 364-374, 出版日 2014年06月18日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - RDFデータの極小モデル推論に基づく記述論理ALCH(D) の概念学習
長井拓馬, 兼岩 憲
ラスト(シニア)オーサー, 人工知能学会論文誌, 29巻, 3号, 掲載ページ 343-355, 出版日 2014年06月05日, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - Event Inference with Relation and Meta-relation Type Hierarchies in Conceptual Structure Theory
Philip H. P. Nguyen and Ken Kaneiwa
Applied Artificial Intelligence, 28巻, 2号, 掲載ページ 139-177, 出版日 2014年03月, 査読付, 国内誌
研究論文(学術雑誌), 英語 - Community-Driven and Ontology-Based Biological Knowledge Management: A Hybrid Approach to Harnessing Collective Intelligence
Michiaki Iwazume and Ken Kaneiwa
14th ACIS International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel & Distributed Computing (SNPD 2013), 掲載ページ 387-393, 出版日 2013年07月01日, 査読付, 国際誌
研究論文(国際会議プロシーディングス), 英語 - An order-sorted query system for sort, predicate, and meta-predicate hierarchies
Ken Kaneiwa; Philip H. P. Nguyen
KNOWLEDGE AND INFORMATION SYSTEMS, SPRINGER LONDON LTD, 35巻, 2号, 掲載ページ 463-492, 出版日 2013年05月, 査読付, This paper presents a decidable order-sorted query system for reasoning between ontologies and rules. We describe order-sorted logic programming with sort, predicate, and meta-predicate hierarchies (OSL3h), which derives predicate and meta-predicate assertions. Meta-level predicates (predicates of predicates) are useful for representing relationships between predicate formulas, and further, they conceptually yield a hierarchy similar to the hierarchies of sorts and predicates. By extending the order-sorted Horn-clause calculus, we develop a query-answering system in OSL3h that can answer queries such as atoms and meta-atoms generalized by containing predicate variables. We show that the expressive query-answering system computes every generalized query in single exponential time, that is, the complexity of our query system is equal to that of DATALOG.
研究論文(学術雑誌), 英語 - Reasoning about Resources and Information: A Linear Logic Approach
Norihiro Kamide and Ken Kaneiwa
Fundamenta Informaticae, 125巻, 1号, 掲載ページ 51-70, 出版日 2013年, 査読付, 国際誌
研究論文(学術雑誌), 英語 - Existential rigidity and many modalities in order-sorted logic
Kaneiwa Ken
Knowledge-Based Systems, Elsevier B.V., 24巻, 5号, 掲載ページ 629-641, 出版日 2011年07月01日, Order-sorted logic is a useful tool for knowledge representation and reasoning because it enables representation of sorted terms and formulas along with partially ordered sorts (called sort-hierarchy). However, this logic cannot represent more complex sorted expressions when they are true in any possible world (as rigid) or some possible worlds (as modality) such as time, space, belief, or situation. In this study, we extend order-sorted logic by introducing existential rigidity and many modalities. In the extended logic, sorted modal formulas are interpreted over the Cartesian product of sets of possible worlds. We present a new labeled tableau calculus to check the (un)satisfiability and validity of sorted modal formulas. - A rough set approach to multiple dataset analysis
Kaneiwa Ken
Applied Soft Computing, Elsevier B.V., 11巻, 2号, 掲載ページ 2538-2547, 出版日 2011年03月01日, In the area of data mining, the discovery of valuable changes and connections (e.g., causality) from multiple data sets has been recognized as an important issue. This issue essentially differs from finding statistical associations in a single data set because it is complicated by the different data behaviors and relationships across multiple data sets. Using rough set theory, this paper proposes a change and connection mining algorithm for discovering a time delay between the quantitative changes in the data of two temporal information systems and for generating the association rules of changes from their connected decision table. We establish evaluation criteria for the connectedness of two temporal information systems with varying time delays by calculating weight-based accuracy and coverage of the association rules of changes, adjusted by a fuzzy membership function. - A Time and Situation Dependent Semantics for Ontological Property Classification
Ken Kaneiwa; Riichiro Mizoguchi
IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS, IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG, E94D巻, 3号, 掲載ページ 639-647, 出版日 2011年03月, 査読付, This paper proposes a new semantics that characterizes the time and/or situation dependencies of properties, together with the ontological notion of existential rigidity. For this purpose, we present order-sorted tempo-situational logic (OSTSL) with rigid/anti-rigid sorts and an existential predicate. In this logic, rigid/anti-rigid sorted terms enable the expressions for sortal properties, and temporal and situational operators suitably represent the ontological axioms of existential rigidity and time and/or situation dependencies. A specific semantics of OSTSL adheres to the temporal and situational behaviors of properties based on existential rigidity. As a result, the semantics guarantees that the ontological axioms of properties expressed by sorted tempo-situational formulas are logically valid.
研究論文(学術雑誌), 英語 - Paraconsistent Computation Tree Logic
Ken Kaneiwa and Norihiro Kamide
New Generation Computing, 29巻, 4号, 掲載ページ 391-408, 出版日 2011年, 査読付, 国際誌
研究論文(学術雑誌), 英語 - セマンティックWeb のためのイベントオントロジー
兼岩 憲, 岩爪道昭
コンピュータソフトウェア, 28巻, 3号, 掲載ページ 153-166, 出版日 2011年, 査読付, 国内誌
研究論文(学術雑誌), 日本語 - On the complexities of consistency checking for restricted UML class diagrams
Ken Kaneiwa; Ken Satoh
THEORETICAL COMPUTER SCIENCE, ELSEVIER SCIENCE BV, 411巻, 2号, 掲載ページ 301-323, 出版日 2010年01月, 査読付, Automatic debugging of UML class diagrams helps in the visual specification of software systems because users cannot detect errors in logical consistency easily. This study focuses on the tractable consistency checking of UML class diagrams. We accurately identify inconsistencies in these diagrams by translating them into first-order predicate logic that is generalized by counting quantifiers and classify their expressivities by eliminating certain components. We introduce optimized algorithms that compute the respective consistencies of class diagrams of different expressive powers in P, NP, PSPACE, or EXPTIME with respect to the size of the class diagrams. In particular, owing to the restrictions imposed on attribute value types, the complexities of consistency checking of class diagrams decrease from EXPTIME to P and PSPACE in two cases: (i) when the class diagrams contain disjointness constraints and overwriting/multiple inheritances and (ii) when the class diagrams contain both these components along with completeness constraints. Additionally, we confirm the existence of a restriction of class diagrams that prevents any logical inconsistency. (C) 2009 Elsevier B.V. All rights reserved.
研究論文(学術雑誌), 英語 - Distributed reasoning with ontologies and rules in order-sorted logic programming
Ken Kaneiwa; Riichiro Mizoguchi
JOURNAL OF WEB SEMANTICS, ELSEVIER SCIENCE BV, 7巻, 3号, 掲載ページ 252-270, 出版日 2009年09月, 査読付, Integrating ontologies and rules on the Semantic Web enables software agents to interoperate between them; however, this leads to two problems. First, reasoning services in SWRL (a combination of OWL and RuleML) are not decidable. Second, no studies have focused on distributed reasoning services for integrating ontologies and rules in multiple knowledge bases. In order to address these problems, we consider distributed reasoning services for ontologies and rules with decidable and effective computation. In this paper, we describe multiple order-sorted logic programming that transfers rigid properties from knowledge bases. Our order-sorted logic contains types (rigid sorts), non-rigid sorts, and unary predicates that distinctly express essential sorts, non-essential sorts, and non-sortal properties. We formalize the order-sorted Horn-clause calculus for such properties in a single knowledge base. This calculus is extended by embedding rigid-property derivation for multiple knowledge bases, each of which can transfer rigid-property information from other knowledge bases. In order to enable the reasoning to be effective and decidable, we design a query-answering system that combines order-sorted linear resolution and rigid-property resolution as top-down algorithms. (C) 2009 Elsevier B.V. All rights reserved.
研究論文(学術雑誌), 英語 - Decidable order-sorted logic programming for ontologies and rules with argument restructuring
Ken Kaneiwa; Philip H.P. Nguyen
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Springer Verlag, 5823巻, 掲載ページ 328-343, 出版日 2009年, This paper presents a decidable fragment for combining ontologies and rules in order-sorted logic programming. We describe order-sorted logic programming with sort, predicate, and meta-predicate hierarchies for deriving predicate and meta-predicate assertions. Meta-level predicates (predicates of predicates) are useful for representing relationships between predicate formulas, and further, they conceptually yield a hierarchy similar to the hierarchies of sorts and predicates. By extending the order-sorted Horn-clause calculus, we develop a query-answering system that can answer queries such as atoms and meta-atoms generalized by containing predicate variables. We show that the expressive query-answering system computes every generalized query in single exponential time, i.e., the complexity of our query system is equal to that of DATALOG. © Springer-Verlag Berlin Heidelberg 2009.
研究論文(国際会議プロシーディングス), 英語 - Contradiction Finding and Minimal Recover for UML class diagrams
Ken Satoh; Ken Kaneiwa; Takeaki Uno
ASE2006, 掲載ページ 277-280, 出版日 2006年09月, 査読付
研究論文(国際会議プロシーディングス), 英語 - Consistency checking algorithms for restricted UML class diagrams
Ken Kaneiwa; Ken Satoh
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3861巻, 掲載ページ 219-239, 出版日 2006年, 査読付, Automatic debugging of UML class diagrams helps in the visual specification of software systems because users cannot detect errors in logical inconsistency easily. This paper focuses on tractable consistency checking of UML class diagrams. We accurately identify inconsistencies in these diagrams by translating them into first-order predicate logic generalized by counting quantifiers and classify their expressivities by eliminating some components. For class diagrams of different expressive powers, we introduce optimized algorithms that compute their respective consistencies in P, NP, PSPACE, or EXPTIME with respect to the size of a class diagram. In particular, for two cases in which class diagrams contain (i) disjointness constraints and overwriting/multiple inheritances and (ii) these components along with completeness constraints, the restriction of attribute value types decreases the complexities from EXPTIME to P and PSPACE. Additionally, we confirm the existence of a meaningful restriction of class diagrams that prevents any logical inconsistency. © Springer-Verlag Berlin Heidelberg 2006.
研究論文(国際会議プロシーディングス), 英語 - Formal Ontology and Extensions of Order-Sorted Logic
Ken Kaneiwa; Riichiro Mizoguchi
Transactions of the Japanese Society for Artificial Intelligence, 20巻, 掲載ページ 387-395, 出版日 2005年, 査読付, This paper presents several extensions of order-sorted logic based on the notion of property classification in formal ontology. The refined order-sorted language contains types (as rigid sorts), anti-rigid sorts, and unary predicates in order to distinctly express the following properties: substantial sorts, non-substantial sorts, and non-sortal properties. For many separated knowledge bases constructed using the logic, we propose an enriched reasoning mechanism such that each independent knowledge base can extract rigid property information from other knowledge bases (called rigid property derivation). Additionally, we classify (i) anti-rigid properties on the basis of their dependences on time, situation, and belief and (ii) non-sortal properties on the basis of the countability of the elements of the properties. keywords: formal ontology, order-sorted logic, ontological property classification, rigidity. © 2005, The Japanese Society for Artificial Intelligence. All rights reserved.
研究論文(学術雑誌), 英語 - An order-sorted quantified modal logic for meta-ontology
Ken Kaneiwa; Riichiro Mizoguchi
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3702巻, 掲載ページ 169-184, 出版日 2005年, 査読付, The notions of meta-ontology enhance the ability to process knowledge in information systems
in particular, ontological property classification deals with the kinds of properties in taxonomic knowledge based on a philosophical analysis. The goal of this paper is to devise a reasoning mechanism to check the ontological and logical consistency of knowledge bases, which is important for reasoning services on taxonornic knowledge. We first consider an ontological property classification that is extended to capture individual existence and time and situation dependencies. To incorporate the notion into logical reasoning, we formalize an order-sorted modal logic that involves rigidity, sortality, and three kinds of modal operators (temporal/situational/any world). The sorted expressions and modalities establish axioms with respect to properties, implying the truth of properties in different kinds of possible worlds and in varying domains in Kripke semantics. We provide a prefixed tableau calculus to test the satisfiability of such sorted modal formulas, which validates the ontological axioms of properties. © Springer-Verlag Berlin Heidelberg 2005.
研究論文(国際会議プロシーディングス), 英語 - Order-sorted logic programming with predicate hierarchy
K Kaneiwa
ARTIFICIAL INTELLIGENCE, ELSEVIER SCIENCE BV, 158巻, 2号, 掲載ページ 155-188, 出版日 2004年10月, 査読付, Order-sorted logic has been formalized as first-order logic with sorted terms where sorts are ordered to build a hierarchy (called a sort-hierarchy). These sorted logics lead to useful expressions and inference methods for structural knowledge that ordinary first-order logic lacks. Nitta et al. pointed out that for legal reasoning a sort-hierarchy (or a sorted term) is not sufficient to describe structural knowledge for event assertions, which express facts caused at some particular time and place. The event assertions are represented by predicates with n arguments (i.e., n-ary predicates), and then a particular kind of hierarchy (called a predicate hierarchy) is built by a relationship among the predicates. To deal with such a predicate hierarchy, which is more intricate than a sort-hierarchy, Nitta et al. implemented a typed (sorted) logic programming language extended to include a hierarchy of verbal concepts (corresponding to predicates). However, the inference system lacks a theoretical foundation because its hierarchical expressions exceed the formalization of order-sorted logic. In this paper, we formalize a logic programming language with not only a sort-hierarchy but also a predicate hierarchy. This language can derive general and concrete expressions in the two kinds of hierarchies. For the hierarchical reasoning of predicates, we propose a manipulation of arguments in which surplus and missing arguments in derived predicates are eliminated and supplemented. As discussed by Allen, McDermott and Shoham in research on temporal logic and as applied by Nitta et al. to legal reasoning, if each predicate is interpreted as an event or action (not as a static property), then missing arguments should be supplemented by existential terms in the argument manipulation. Based on this, we develop a Horn clause resolution system extended to add inference rules of predicate hierarchies. With a semantic model restricted by interpreting a predicate hierarchy, the soundness and completeness of the Horn-clause resolution is proven. (C) 2004 Elsevier B.V. All rights reserved.
研究論文(学術雑誌), 英語 - オントロジー構築に対する概念階層の整合性
兼岩 憲; Ken Kaneiwa
コンピュータソフトウェア = Computer software, Japan Society for Software Science and Technology, 21巻, 2号, 掲載ページ 94-106, 出版日 2004年03月25日, オントロジーは一般的に概念階層によって設計されるが,その概念階層の整合性を判断するには,概念階層を形式的に定義する必要がある.Description Logicやオーダーソート論理をはじめとする論理型言語は,概念階層のための表現を備えており,オントロジー記述言語に理論的な基盤を与えることができる.しかしそこで定義される概念階層は,個体の集合を示したクラス表現の階層に限られている.本論文では,実際のオントロジー設計において様々な自然言語の語彙で概念階層が構築されることを想定して,クラス表現に限定しない,ソート,定数,関数および述語の非論理記号が混ざり合った概念階層を形式化する.そのうえで,この複雑な概念階層を安全に構築する方法を提案する.
日本語 - Occurrence Logic with Temporal Heredity.
Suguru Yoshioka; Ken Kaneiwa; Satoshi Tojo
Proceedings of the 1st Indian International Conference on Artificial Intelligence, IICAI 2003, Hyderabad, India, December 18-20, 2003, IICAI, 掲載ページ 1296-1309, 出版日 2003年, 査読付
研究論文(国際会議プロシーディングス) - An Order-Sorted Resolution with Implicitly Negative Sorts.
Ken Kaneiwa; Satoshi Tojo
Logic Programming, 17th International Conference, ICLP 2001, Paphos, Cyprus, November 26 - December 1, 2001, Proceedings, Springer, 2237巻, 掲載ページ 300-314, 出版日 2001年, 査読付, We usually use natural language vocabulary for sort names in order-sorted logics, and some sort names may contradict other sort names in the sort-hierarchy. These implicit negations, called lexical negations in linguistics, are not explicitly prefixed by the negation connective. In this paper, we propose the notions of structured sorts, sort relations, and the contradiction in the sort-hierarchy. These notions specify the properties of these implicit negations and the classical negation, and thus, we can declare the exclusivity and the totality between two sorts, one of which is affirmative while the other is negative. We regard the negative affix as a strong negation operator, and the negative lexicon as an antonymous sort that is exclusive to its counterpart in the hierarchy. In order to infer from these negations, we integrate a structured sort constraint system into a clausal inference system.
研究論文(国際会議プロシーディングス) - Event and property in order-sorted logic
Ken Kaneiwa; Satoshi Tojo
Computer Software, 17巻, 2号, 掲載ページ 10-24, 出版日 2000年
研究論文(学術雑誌)
MISC
- DL : Description Logics(「哲学とAIにおける対象世界モデリング」〔第6回〕)
兼岩 憲; 佐藤 健; Ken Kaneiwa; Ken Satoh; Foundations of Informatics Research Division National Institute of Informatics; Foundations of Informatics Research Division National Institute of Informatics
人工知能学会, 出版日 2003年01月01日, 人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence, 18巻, 1号, 掲載ページ 73-82, 日本語, 0912-8085, 110002806486, AN10067140
講演・口頭発表等
- グラフニューラルネットワークに対する同種性を用いた母集団グラフの効率化
吉永 康希; 兼岩 憲
口頭発表(一般), 第133回人工知能学会研究会資料 知識ベースシステム研究会
発表日 2024年12月05日 - オントロジーの論理構造と語彙情報を考慮した概念包含のための埋め込み学習
飯島 彩友; 兼岩 憲
口頭発表(一般), 第62回セマンティックウェブとオントロジー研究会
発表日 2024年03月23日 - 確率的除去を組み込んだ検索拡張された言語モデルによる帰納的関係予測
岩本 有生; 兼岩 憲
口頭発表(一般), 日本語, 第125回人工知能基本問題研究会
発表日 2023年08月29日 - 概念包含の論理的推論に対するロール埋め込み手法
山中 佑斗, 白石 幸寛, 兼岩 憲
口頭発表(一般), 日本語, 第127回 知識ベースシステム研究会 (SIG-KBS)
発表日 2022年11月23日
開催期間 2022年11月23日 - グラフ構造のIGR多重特徴化によるグラフニューラルネットワーク
大石 悠河, 兼岩 憲
口頭発表(一般), 日本語, 第127回 知識ベースシステム研究会 (SIG-KBS)
発表日 2022年11月23日
開催期間 2022年11月23日 - 大規模RDFグラフに対する効率的なクエリ解決と高圧縮なインデックス化
藤原 浩司, 兼岩 憲
口頭発表(一般), 日本語, 電子情報通信学会 人工知能と知識処理研究会(AI)
発表日 2022年07月04日
開催期間 2022年07月04日 - Decidable Order-Sorted Logic Programming for Ontologies and Rules with Argument Restructuring
Ken Kaneiwa; Philip H; P; Nguyen
8th International Semantic Web Conference (ISWC 2009) (acceptance rate: 17.3%)., 査読付
発表日 2009年10月
開催期間 2009年10月
担当経験のある科目_授業
所属学協会
共同研究・競争的資金等の研究課題
- セマンティックWebデータの多様性に対する推論と学習の基盤技術
兼岩 憲
日本学術振興会, 科学研究費助成事業 基盤研究(C), 電気通信大学, 基盤研究(C), 2021年度は、RDFデータやオントロジーによる2つの異なるレベルのセマンティックWebデータを学習して、リンク関係、論理的関係、およびクラス属性を推定する学習型推論システムを開発した。
リンクトデータを表現するRDFデータ(主語、述語、目的語からなるトリプル)に対する埋め込み学習とともに、オントロジーを表現する記述論理の知識ベースに対する埋め込み学習を実装した。これにより、セマンティックWebデータから記述論理の推論アルゴリズムを使って論理的に推論するだけでなく、訓練済みの学習器から実データに不足している知識を新たに推定できるようになっている。即ち、従来の推論システムでは論理的な帰結しか導けないのに対して、学習型推論システムでは知識ベースに含まれていない概念のインスタンスデータ、個体間の二項関係、概念間の包含関係などを導くことができる。今回、RDFデータとオントロジーの両タイプに機械学習を適用することで、ドメインオントロジーをリンクトデータに融合させた知識ベースでの推論が期待できる。
さらに、前年度までに開発した特徴抽出の手法により、リンクトデータと類似性の高いグラフデータベースに対して、不要な知識をスキップしてデータを特徴ベクトル化する手法を適用した。これは人間関係や科学論文関係などのようなネットワーク構造で表されるグラフデータにおいて、各対象物(人間や科学論文)がどのようなクラスに属すか学習する。本分野で用いられているグラフニューラルネットワーク(GNN)と本研究が提案している特徴抽出の手法を組み合わせて高い推定能力(正解率)を実現した。, 18K11547
研究期間 2018年04月 - 2023年03月 - セマンティックWebデータに対する推論基盤技術の研究開発
兼岩 憲
日本学術振興会, 科学研究費助成事業 基盤研究(C), 電気通信大学, 基盤研究(C), 本研究では、RDFデータから記述論理のクエリに答える意味的推論システムを開発した。この意味的推論システムとともに、大規模データを推論・検索するためのRDFデータストアを独自に開発した。さらに、ユーザーが簡単に検索できるように、自然言語文による質問とキーワードによる検索方法を実装している。その際、ユーザーの漠然とした関心から検索や質問に答えるために、キーワードや自然言語文からの推論に加えて、機械学習やデータマイニングによってユーザーが求める知識を推定する方法を提案した。, 25330360
研究期間 2013年04月01日 - 2018年03月31日 - サイバー・フィジカルシステムのためのスケーラブルなセマンティックサービス基盤技術
岩爪 道昭; 兼岩 憲; 小林 一郎; 小林 一郎
日本学術振興会, 科学研究費助成事業 基盤研究(C), 独立行政法人情報通信研究機構, 基盤研究(C), 近年急速に流通しているOpen Linked Dataのためのサービス基盤技術の研究開発に取り組んだ。(1)イベンドデータのモデリング手法を開発し、これをDBPediaなどの意味的資源に適用し、プロパティ情報に着目した定式化を行った。(2)(1)で定式化した意味的プロパティ情報のランキング手法を定式化・実装し、DBPediaを対象にした実証実験を行い、その有効性について確認した。(3)Open Linked Data等の意味的資源を収集・蓄積・解析するためのクローラー、分散データストア等のミドルウェアおよびそれを支える大型計算機システムの動態情報を可視化を試作し、動作検証を行った。, 23500199
研究期間 2011年04月28日 - 2015年03月31日 - 決定ルールと構造オントロジーによる意思決定知識ベースシステム
兼岩 憲
日本学術振興会, 科学研究費助成事業 若手研究(B), 若手研究(B), 研究成果の概要(和文):意思決定ルールと集合構造オントロジーを抽出するために、論理的推論とラフ集合データマイニングの異なる2つのタイプの推論メカニズムを研究開発した。それにより、データ分析と事例ベース推論の2つの意思決定の側面を取り入れた知識ベースシステムの基盤を与えている。さらに、実際の知識構造データから意思決定支援へ結びつけるために、ユーザーの入力を起点にして、そこからルールとオントロジーの推論を適用する方法を提案した。, 20700147
研究期間 2008年 - 2011年 - パスウェイ知識表現における矛盾検証の理論的解析と技術開発
福田 賢一郎; 兼岩 憲
日本学術振興会, 科学研究費助成事業 特定領域研究, 独立行政法人産業技術総合研究所, 特定領域研究, 疾病や生物プロセスなど,生命現象のメカニズムを対象とした知識処理では,論文やデータベースのコメント欄に自然言語で記述された知識の電子化手法が従来の課題であった。そのため遺伝子機能を記述するためのオントロジーや,論文から該当知識を抽出するテキストマイニングなどの自然言語処理技術の研究開発が世界的に進められてきた。また,生命現象の分子メカニズムを記述するパスウェイデータベースについては,オントロジーで詳細に意味を定義したデータフォーマットを標準化させようという国際的な取り組みがすすんでいる。これらの成果により,近い将来,専門家によってキュレートされた知識自然言語処理システムが自動抽出した知識,そして様々な実験技術から導出された知識というように精度も性質も異なる情報が共通のプラットホーム上で入手可能になる事が期待される。このような状況を見据えて,今後は互いに整合性を持っていると保障されていない異なる出自のデータを対象とした知識処理技術の開発が必要となる。そこで本研究ではパスウェイデータの矛盾検出に関する知識表現の基礎研究を展開し,本年度は主として以下の成果を得ることができた。1)イベント知識ベース構築の為の形式オントロジーの開発。2)BioPAX オントロジーで記述されたデータをバックエンドデータベースに格納する為の検索ライブラリpaxtoolsの開発。3)INOHデータのBioPAXフォーマットでのリリース。1)はイベント表現を蓄積する知識ベースの基盤となる上位オントロジーである。パスウェイは様々な要素が多様な関係性をもちながら細胞機能という現象を引き起こす分子メカニズムであるが,この現象の表現を一般化したものがイベントである。, 18016032
研究期間 2006年 - 2008年 - 形式オントロジーと矛盾要因に関する研究
兼岩 憲
日本学術振興会, 科学研究費助成事業 若手研究(B), 若手研究(B), 最終年度では,これまでの2年間で検討した矛盾要因(論理的な矛盾とオントロジー的な矛盾)の提案および形式化したオントロジーの論理型言語の実用性を検討した。特に,実例による検証と数学的な証明による理論的な検証を行った。実証には,実世界で重要となるイベント知識を含んだ知識表現や推論の有用性を検証するために,法律分野(法的推論)への応用を検討した。 本年度のアカデミックな研究成果として,法的推論システムの実現のために法令文のルール知識に加えて判例を描写するイベント(一回性の事象)のための表現手段を提案した。法律判例の中身は複数のイベントによって展開されており,動的な知識表現に対する矛盾とその論理的な推論を法的推論システムに取り込まなければならない。イベントは一回性・一時性を持った動作あるいはアクションであり,静的なプロパティと対比される概念である。本研究では,法的推論で現れる語彙や記述を例に用い,イベントの量化,ソート階層およびイベント間の合成と排他性を導入した知識表現とその論理型言語(イベント論理と呼ぶ)を実現した。この言語は,イベントを定数,ソート,述語および変数として見なし,イベント言明のための知識表現と推論を可能にする。さらに,法的推論システムを実現する推論メカニズムの基盤を与えるために,イベント論理に対するソート付きのタブロー計算を設計して,その反駁推論による質問応答メカニズムを設計した。, 17700164
研究期間 2005年 - 2007年 - 宣言的プログラミングにおけるソフトウェア発展の研究
佐藤 健; 宇野 毅明; 兼岩 憲
日本学術振興会, 科学研究費助成事業 特定領域研究, 国立情報学研究所, 特定領域研究, 現在のように、ソフトウェアのバージョンアップが頻繁に生じる状況では、ソフトウェア変更管理(ソフトウェア発展)の技術は大変重要なものである。本研究では、宣言的プログラミングにおけるソフトウェア発展を考える。これは、論理型プログラミングや関数型プログラミングなどの宣言的なプログラミングが数学的な体系に基づいているため、ソフトウェア発展技術の構築のための理論が構築しやすいと考えたためである。 本年度は、前年度行ったUML(Unified Modelling Language)クラス図の矛盾検出および自動修正の検討を発展させ、論理プログラミングによる実装を行った。クラス図の各コンポーネントを論理プログラミングのルールに変換し、そのルール上での矛盾の導出がクラス図での矛盾検出と同値になるような手法の開発を行った。さらに論理プログラミングにおけるメタルールを用いることでルール集合のどの部分が矛盾であるかを検出できるようになった。このルール集合の矛盾部分に対して極小被覆集合アルゴリズムを用いて極小にUMLの部分を消去することで矛盾除去ができる手法を開発した。この手法を検証するためにプロトタイプシステムを作成し、高速処理への見通しを得た。 また、UMLクラス図の部分クラスにおいて矛盾検証そのものが不要になるものを発見し、さらに部分クラスを分類して各表現能力に応じて矛盾検証アルゴリズムがどの程度計算量的に複雑になるかについても見当した。, 16016284
研究期間 2004年 - 2005年 - オントロジー指向による論理体系の研究
兼岩 憲
日本学術振興会, 科学研究費助成事業 若手研究(B), 国立情報学研究所, 若手研究(B), 昨年度は、オントロジーを構築する概念階層の多様性について分析し、多様性から生じる概念記述とその推論の難しさを扱う方法論を提案した。平成16年度は、形式オントロジーで提案されている属性分類の考え方を順序ソート論理に取り入れ、それによって拡張される知識ベース推論の枠組みを提案した。研究成果の詳細は、以下の通りである。 順序ソート論理においてソートと単項述語は意味論的に同等と見なされるが、形式オントロジーではそれらは異なった種類の属性として分類される(例えば、ソートと非ソートの区別)。こうしたオントロジー的な概念分析は、知識表現と推論の中で個体の属性(またはソート)を扱うのに有用な指針を与えてくれる。本研究では、形式オントロジーの属性分類を取り込むことで順序ソート論理の構文、意味論および推論システムを拡張する方法を提案した。この拡張順序ソート論理は、本質ソート、非本質ソートおよび非ソート属性に分類される概念を、タイプ(rigidソート)、non-rigidソートおよび単項述語で表した上で、インスタンス言明と包摂関係のrigidityを考慮して形式化される。さらに本質属性が状況などに依存しないで永久的に成り立つ性質を利用して、独立した複数の知識ベースを想定したとき、各知識ベースが(共通の知識として)rigidな属性情報を他の知識ベースから抽出できる推論メカニズム(本質属性の導出と呼ぶ)を設計した。, 14780311
研究期間 2002年 - 2004年