著者

小長谷 明彦 Akihiko Konagaya 北陸先端科学技術大学院大学知識科学研究科 School of Knowledge Science Japan Advanced Insititute of Science and Technology Hokuriku

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.13, no.1, pp.38-40, 1998-01-01

著者

小林 哲郎

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.27, no.1, pp.51-58, 2012-01-01

著者

山口 喜教 樋口 哲也 坂井 修一 Yoshinori Yamaguchi Tetsuya Higuchi Shuichi Sakai 電子技術総合研究所電子計算機部 電子技術総合研究所電子計算機部 電子技術総合研究所電子計算機部 Computer Systems Division Electrotechnical Laboratory Computer Systems Division Electrotechnical Laboratory Computer Systems Division Electrotechnical Laboratory

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.2, no.4, pp.459-471, 1987-12-01

著者

相良 毅 井口 誠 藤本 和則 Takeshi Sagara Makoto Iguchi Kazunori Fujimoto 東京大学生産技術研究所 フランステレコム株式会社 有限会社フジモト・リサーチパーク Institute of Industrial Science The University of Tokyo Research & Development Tokyo France Telecom Research & Development Tokyo Artificial Intelligence Laboratory Fujimoto Research Park Co. LTD.

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.21, no.4, pp.430-437, 2006-07-01 (Released:2017-02-17)

著者

武田 英明 冨山 哲男 吉川 弘之

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.7, no.5, pp.877-887, 1992-09-01

被引用文献数

15

---

This paper presents a computable design process model that is based on a logical framework. A computable design process model is useful not only to investigate designing activities but also to realize intelligent CAD (Computer-Aided-Design) systems. We propose a computable model based on a non-classical logical framework including abduction, circumscription, meta-level reasoning, and modal logic. First, we discuss design processes with a cognitive model of design processes. We point out four problems that are difficult to solve in the classical logical framework. Second, we formalize design processes in a non-classical logical framework. In this formalization, a design process is regarded as an iterative process of abduction, deduction, and circumscription. Abduction is used to obtain candidate descriptions of the design object, deduction to find out properties of the design object, and circumscription to solve inconsistency occurred in the design process. These types of inference are used for knowledge about objects ; on the other hand meta-level reasoning based on knowledge about action is used to reason out what should be done next. Furthermore we introduce data semantics to represent transitions of design states. Data semantics is a kind of modal logic which allows not only truth value t and f but also u (unknown), and a design process is interpreted as a process generating possible worlds sequentially. Then, we illustrate a design simulator based on this logical model. We discuss formalization and implementation of abductive inference and utilization of other inferences that are needed to implement a design simulator. We show that it can trace a design process in which design objects are gradually refined as the design proceeds.

著者

鈴木 優

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.26, no.1, 2011-01-01

著者

石川 真澄

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.5, no.5, pp.595-603, 1990-09-01

被引用文献数

104

---

Learning in connectionist models has two aspects : the first aspect being the reproduction of the mapping from input to output patterns, and the second being the discovery of regularity in these training patterns. The backpropagation learning algorithm stresses the former aspect, as can be seen from its criterion function of the sum of squared output errors. The present paper, on the other hand, lays emphasis on the latter aspect. In the backpropagation learning of feedforward type models it is of nesessity to determine, beforehand, the number of layers and the number of hidden units in each layer. Since this prior determination is, in general, difficult, a trial and error procedure is inevitable, which is quite time consuming. To overcome this difficulty and to generate a small sized network, the present paper proposes a learning algorithm with forgetting of link weights. This forgetting is realized by adding the sum of absolute values of link weights to the criterion in the backpropagation algorithm. This algorithm generates a skeletal structure, in which the numbers of links and units used are kept as small as possible. As by-products of this algorithm it has various advantages : ease of interpretation of hidden units and improved generalization power of the resulting models. This algorithm alone causes the following two difficulties : emergence of distributed representation on hidden layers, which makes the interpretation of hidden units difficult, and a poor criterion value after learning due to the added term in the criterion function. To resolve these difficulties a structural learning algorithm is proposed, which consists of a series of algorithms : the learning algorithm with forgetting, a hidden units clarification algorithm, and a learning algorithm with selective forgetting. This structural learning algorithm is applied to a problem of discovering a logical function from a given set of pairs of input and output logical values. It is well demonstrated that the resulting skeletal network represents logical structure of the given problem. On the contrary the backpropagation algorithm generates a network far from being skeletal, making the interpretation of hidden units quite difficult. This algorithm is applied to another problem of classifying iris data by Fischer.Generalization power of the structural learning algorithm and that of the backpropagation algorithm are compared. The result of the comparison clearly demonstrates that the former has greater generalization power than the latter.

著者

藤堂 清 松本 俊二 佐藤 智昭

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.5, no.2, pp.213-219, 1990-03-01

被引用文献数

3

---

We are developing an intelligent development environment for knowledge systems based on ES/SDEM (Software Development Engineering Methodology for Expert Systems). The environment provides not only development tools such as editors or debugger, but also knowledge base called Knowledge Ware (KW) which can be used on different applications. KW is written using an object-oriented language on Common Lisp, and will provide the following advantages. ・User can build up expert systems rapidly using differential programming technique. ・Adding the parts made for an expert system to KW, it is easy to use the same parts for developing another expert systems. In this paper, the approach for building KW and the function of KW will be shown.

著者

松崎 有理

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.27, no.5, pp.528-529, 2012-09-01

被引用文献数

1

著者

矢入 郁子 原井 洋明 村田 正幸

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.24, no.3, pp.416-423, 2009-05-01

被引用文献数

1

著者

岡谷 貴之

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.28, no.6, pp.962-974, 2013-11-01

著者

大塚 真吾 喜連川 優

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.21, no.4, pp.410-415, 2006-07-01

参考文献数

23

被引用文献数

6 4

著者

秋葉 友良

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.17, no.5, pp.553-558, 2002-09-01

被引用文献数

4

著者

山村 雅幸 小野 貴久 小林 重信

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.7, no.6, pp.1049-1059, 1992-11-01

被引用文献数

105

---

Genetic Algorithms (GA) is a new learning paradigm that models a natural evolution mechanism. The framework of GA straightly corresponds to an optimization problem. They are classified into functional optimization and combinatorial one, and have been studied in different manners. GA can be applied to both types of problems and moreover their combinations. According to generations, GA will discover and accumulate building blocks in the form of schemata, and find the global solution as their combinations. It is said GA can find the global solution rapidly if the population holds sufficient varieties. However, this expectation has not been confirmed rigidly. Indeed, there are some problems pointed out such as the early convergence problem in functional optimization, and the encode/decode-crossover problem in combinatorial one. In this paper, we give a solution to the encode/decode-crossover problem for traveling salesman problems (TSP) with a character-preserving GA. In section 2, we define the encode/decode-crossover problem. The encode-decode problem is to define a correspondence between GA space and problem space. The crossover problem is to define a crossover method in GA space. They are closely related to the performance of GA. We point out some problems with conventional approaches for TSP. We propose three criteria to define better encode/decode ; the completeness, soundness and non-redundancy. We also propose a criterion to define better crossover ; character-preservingness. In section 3, we propose a character-preserving GA. In TSP, good subtours are worth preserving for descendants. We propose a subtour exchange crossover, that will not break subtours as possible. We also propose a compress method to improve efficiency. In section 4, we design an experiment to confirm usefulness of our character-preserving GA. We use a double-circled TSP in which the same numbers of cities are placed on two concentrated circles. There are two kinds of local solutions ; "C"-type and "O"-type. The ratio between outer and inner radius determines which is the optimum solution. We vary the radius ratio and see how much optimal solutions are obtained. In the result, character-preserving GA finds optimal solutions effectively.

著者

神島 敏弘

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.18, no.2, pp.170-176, 2003-03-01

被引用文献数

12 7

著者

林 晋 永井 和 宮崎 泉

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.25, no.1, pp.24-31, 2010-01-01

被引用文献数

1

著者

松村 真宏

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.26, no.5, pp.425-431, 2011-09-01

被引用文献数

1

著者

鷲尾 隆 Takashi Washio 大阪大学産業科学研究所知能システム科学研究部門 The Institute for Scientific and Industrial Research Osaka University

出版者

人工知能学会

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.23, no.3, pp.362-366, 2008-05-01

参考文献数

2

著者

柴田 博仁

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.19, no.1, 2004-01-01

---

本研究では,文章作成支援という問題を取り上げる.特に,論文や報告書,小説,エッセイなどの作成で見られるように,ある問題やテーマに対して関連する情報の収集と独自のアイディアが求められ,最終的には表現や構成を吟味したうえで文章としてまとめる活動を支援することを目的とする.このような文章作成は,問題が明確に定まらず,実践の中で問題を発見し,解決していくデザインプロセスの一種として考えることができる.また,その過程で生じる問題を解決するためには,これまでの概念空間から別の空間へとジャンプする創造的思考が求められる.本研究では,文章作成を「創造的デザイン」として捉え,これまでの創造的思考,デザインプロセス,文章作成に関する研究の成果をもとに,支援の方法論を探る.このような考えに基づき,文章作成に関わるさまざまな活動を一貫して支援する枠組みを提案し,統合的文章作成支援環境iWareを構築した.iWareは,おのおの独自の枠組みに基づいて構築した二つのサブシステムからなる.一つは文章の素材を集め,アイディアの生成を支援するiBox+であり,他方は収集,生成した素材やアイディアを整理し,文章にまとめあげることを支援するiWeaverである.実践的な利用の中で,本研究の枠組みは,日常的な情報収集活動と文章を構築する活動とをシームレスに支援し,文章の構造が定まらない試行錯誤的な状態から明確に構造化されるまでのスムーズな移行を可能とすることを示す.

著者

折原 良平

出版者

社団法人人工知能学会

雑誌

人工知能学会誌 (ISSN:09128085)

巻号頁・発行日

vol.9, no.2, pp.248-257, 1994-03-01

参考文献数

18

被引用文献数

10

---

A system for creativity support "Chie-no-izumi" is presented. Implementing a model of creation based on analogical reasoning, Chie-no-izumi gives users some new concepts which are hints to novel ideas. The framework of analogical reasoning which is employed is "paraphrasing-based analogical reasoning (PA)". 0ne Concept, which is defined in the knowledge base, may be used to define the other concepts. PA generates new concepts by analogically reasoning on such a hierarchic structure of concepts. Domain division, which is postponed in past studies of analogical reasoning, plays an important role in PA. PA can derive several results of analogical reasoning from one knowledge base through different domain divisions. The prototype system for Chie-no-izumi has an interface to natural language (Japanese) inputs, does domain division for analogical reasoning, and graphically exhibits the process of analogical reasoning. Utilizing the information obtained from kana-kanji conversion, the prototype easily parses several kinds of sentences. Introducing heuristics to select external tokens, domain division is automated. Employing not only visual but also audio aides, a user can understand the process of analogical reasoning without difficulty. This paper describes what the model of creation is, how the prototype system works, and how it supports human creativity in planning new products.