Conference Activities & Talks -

オントロジーを利用した医療文献検索システム

平松宙祥, 内藤祐喜, 瀧 寛和, 松田憲幸, 中島二郎, 中村 正, 今川彰久, 松澤佑次

電気学会 産業システム情報化研究会 IIS-03-21  2003.03   The Japan Society of Mechanical Engineers

　View Summary

この研究発表は、医療分野（生活習慣病）のオントロジーを利用することで、従来困難であった意味検索を実現した点に新規性があります。検索キーワードをオントロジーを利用して、拡張語を選定することで、検索者の熟知していない医療分野の新専門用語や、キーワードの一般用語、関連語も考慮した検索を実現しています。専門医（大学病院）、開業医、患者間の医療情報の治療知識ネットワークに実装し、実用性が確認されています。

Design of a tutoring system for an autistic child to learn to fit into society

OGAWA Hisashi, MATSUDA Noriyuki, TAKI Hirokazu, ABE Norihiro

IEICE technical report. Education technology  2002.10.11   The Institute of Electronics, Information and Communication Engineers

　View Summary

Recently, digital educational contents for an autistic person become popular in education at school. However, the digital contents have some problems in education. This paper analyzes the problems to design an authoring tool for the digital contents.

果樹育成の為のレトロフィットエージェント

田尻久幸, 内尾文隆, 張 勇, 松田憲幸, 瀧 寛和

電気学会 産業システム情報化研究会 IIS-02-4  2002.03   The Japan Society of Mechanical Engineers

知的情報検索と文書の構造化を利用したスライドプレゼンテーション

水谷俊介, 松田憲幸, 瀧 寛和, 安部憲広, 堀 聡

電気学会 産業システム情報化研究会 IIS-02-9  2002.03   The Japan Society of Mechanical Engineers

The Model of Prediction for Other Agents' Behavior and Intention using reflection

TAKI Hirokazu, MATSUDA Noriyuki, FUKUI Tohoru, ABE Norihiro, HORI Satoshi

Technical report of IEICE. KBSE  2002.01.17   The Institute of Electronics, Information and Communication Engineers

　View Summary

This paper describes an architecture of a conscious agent using reflection functions. We consider the agent using a production system for inference. In the system, the reflection function observes its working memory. The model of conscious is the relation between "working memory state" and "agent behavior goal". We discuss the conscious model, learning mechanism of the model, and goal selection problem. The prediction problem for other agent intention is also explained.

テキストマイニングによる果樹育成知識の体系化

張 勇, 瀧 寛和, 松田 憲幸, 内尾 文隆, 安部 憲広, 堀 聡, 亀岡 孝治, 井口 信和

農業機械学会誌  2002   The Japanese Society of Agricultural Machinery and Food Engineers

果樹栽培のための知的センサネットワーク

田尻 久幸, 内尾 文隆, 張 勇, 松田 憲幸, 瀧 寛和, 井口 信和, 亀岡 孝治

農業機械学会誌  2002   The Japanese Society of Agricultural Machinery and Food Engineers

調理の記録とレシピの自動対応付けツールの開発 (テーマ:「一般」および「情報の可視化と知識マネジメント」)

白藤 純, 松田 憲幸, 瀧 寛和

知的教育システム研究会  2001.09.29   人工知能学会

インフォメーションマッチメーキングサーバー (特集:「インテリジェントWeb」および一般)

林田 高志, 瀧 寛和, 松田 憲幸

知識ベ-スシステム研究会  2001.09.14   人工知能学会

アクティブWEB:動作表現の伝達 (特集:「インテリジェントWeb」および一般)

頼光 正典, 松田 憲幸, 瀧 寛和

知識ベ-スシステム研究会  2001.09.14   人工知能学会

Simulation based Reasoning for planning expert systems

張 勇, 瀧 寛和, 松田 憲幸, 堀 聡, 安部 憲広

Proc. of Workshop on Web Intelligence and Software Engineering, SIG-KBSE, IEICE  2001.08.24   The Institute of Electronics, Information and Communication Engineers

　View Summary

This paper describes a planning system architecture using simulation based reasoning. The planning system makes an operation sequence from the starting situation to the goal. In case of incomplete knowledge about the environment or scheduled objects, the system always cannot plan the complete solution. It is not easy to know the prediction of behaviors of all objects to be controlled and the environment. The technology for the dynamic planning is reactive planning. The system using this technology can make plants for non-static environment. Even if the system uses this, it is difficult to make plans under incomplete knowledge for the objects in the environment. To predict behaviors of objects, our system uses the simulator of the objects. Our system uses the simulator in the inference of planning. We introduce this architecture and the usage of simulation based reasoning.

オーサリングツールを利用した教育・学習ログの活用の構想 (特集 HAI(Human-Agent Interaction)および一般発表) -- (一般セッション)

松田 憲幸, 白藤 純, 瀧 寛和

知識ベ-スシステム研究会  2001.07.05   人工知能学会

インフォメーションマッチメーキングサーバーによる情報流通基盤

林田高志, 松田憲幸, 瀧 寛和

電気学会 産業システム情報化研究会 IIS-01-5  2001.01.19   The Japan Society of Mechanical Engineers

208 Representation and Management of Design Constraints for Design Document Reuse

TAKI Hirokazu, MATSUDA Noriyuki, YOSHIOKA Katsuhiro, HORI Satoshi, ABE Norihiro

設計シンポジウム講演論文集  2001   "日本機械学会, 精密工学会"

　View Summary

This paper describes the reference and access method for design constraints in the design document. The purpose of the legacy design document represents minimal information of design. Using current information system, the design document can include several kinds of design constraints of reuse and redesign. It is also able to contain recycle constraints, manufacturing constraints, costs information, physical constraints and the designer's will. The system supplies these explicit constraints for redesign and visualizes them for the user. In this paper, we introduce two points that includes parts view and job oriented view of constraints.

拡張現実感による全方位画像の為の情報提供システム (テーマ 「オントロジー活用の実際、産業応用」および一般)

吉岡 克浩, 松田 憲幸, 瀧 寛和

知識ベ-スシステム研究会  2000.11.17   人工知能学会

論文の構成をチェックする手法 (テーマ 「オントロジー活用の実際、産業応用」および一般)

松田 匡弘, 松田 憲幸, 瀧 寛和

知識ベ-スシステム研究会  2000.11.17   人工知能学会

An Aproach of Robot Behavior Learning using Case-based Reasoning

YORIMITSU Masanori, MATSUDA Noriyuki, TAKI Hirokazu, ABE Norihiro

IPSJ SIG Notes. ICS  2000.07.18   Information Processing Society of Japan (IPSJ)

　View Summary

This paper described an application method of case-based reasoning for a robot action learning in real-time environment. The robot action learning is a typical problem of behaviorbased AI. Reinforcement learning, GA and induction are adopted for this problem. On the other hand, the case-based reasoning is suitable for real-time action learning, because it is not necessary for the method to process row acquired cases. The efficiency of the knowledge extracted by this method is not better than the reinforcement learning. But the latter method needs the prior knowledge, and the former one has the advantage that it doesn't need that knowledge. We discuss the application methodology of case-based robot action learning in view of "generalization ratio of cases", "covering ratio of the problem space with cases" and "the number of cases". We use the Robocup simulator for evaluation of this method.

An Aproach of Robot Behavior Learning using Case-based Reasoning

YORIMITSU Masanori, MATSUDA Noriyuki, TAKI Hirokazu, ABE Norihiro

IEICE technical report. Artificial intelligence and knowledge-based processing  2000.07.12   The Institute of Electronics, Information and Communication Engineers

　View Summary

This paper described an application method of case-based reasoning for a robot action learning in real-time environment.The robot action learning is a typical problem of behaviorbased AI.Reinforceinent learning, GA and induction are adopted for this problem.On the other hand, the case-based reasoning is suitable for real-time action learning, bacause it is not necessary for the method to process row acquired cases.The efficiency of the knowledge extracted by this method is not better than the reinforcement learning.But the latter method needs the prior knowledge, and the former one has the advantage that it doesn't need that knowledge.We discuss the application methodology of case-based robot action learning in view of "generalization ratio of cases", "covering ratio of the problem space with cases"and "the number of cases".We use the Robocup simulator for evaluation of this method.

A Support Facility for Browsing in Hypertext based on User's Context : Implementation on WWW

NOMOTO Toyohiro, MATSUDA Noriyuki, HIRASHIMA Tsukasa, TOYODA Jun'ichi

人工知能学会全国大会論文集 = Proceedings of the Annual Conference of JSAI  1998.06.16  

A Framework for Supporting Self-explanations in Collaborative Learning Environment

UI Shinsuke, KASHIHARA Akihiro, MATSUDA Noriyuki, HIRASHIMA Tsukasa, TOYODA Jyun'ichi

教育システム情報学会研究報告  1998.01.24   教育システム情報学会