Lecture Notes in Artificial Intelligence Edited by J. G. Carbonell and J. Siekmann
Subseries of Lecture Notes in Computer Science
4509
Ziad Kobti Dan Wu (Eds.)
Advances in Artificial Intelligence 20th Conference of the Canadian Society for Computational Studies of Intelligence, Canadian AI 2007 Montreal, Canada, May 28-30, 2007 Proceedings
13
Series Editors Jaime G. Carbonell, Carnegie Mellon University, Pittsburgh, PA, USA Jörg Siekmann, University of Saarland, Saarbrücken, Germany Volume Editors Ziad Kobti Dan Wu University of Windsor School of Computer Science 401 Sunset Avenue, Windsor, Ontario, N9B 3P4, Canada E-mail: {kobti,danwu}@uwindsor.ca
Library of Congress Control Number: 2007926815
CR Subject Classification (1998): I.2 LNCS Sublibrary: SL 7 – Artificial Intelligence ISSN ISBN-10 ISBN-13
0302-9743 3-540-72664-0 Springer Berlin Heidelberg New York 978-3-540-72664-7 Springer Berlin Heidelberg New York
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springer.com © Springer-Verlag Berlin Heidelberg 2007 Printed in Germany Typesetting: Camera-ready by author, data conversion by Scientific Publishing Services, Chennai, India Printed on acid-free paper SPIN: 12066940 06/3180 543210
Preface
This volume contains the papers presented at AI 2007, the 20th conference of the Canadian Society for the Computational Study of Intelligence (CSCSI). AI 2007 attracted a new record of 260 paper submissions. Each paper was assigned to three reviewers who tirelessly worked to provide high-quality reviews. Out of these, 46 high-quality papers were accepted for publication by the Program Committee. The organization of AI 2007 has benefited from the collaboration of many individuals. Foremost, we express our appreciation to the Program Committee members and the additional reviewers who provided thorough and timely reviews. We are grateful to Andrei Voronkov and the support team assisting with the EasyChair Conference System that hosted the AI 2007 paper submission and review process. We also extend our thanks to the School of Computer Science at the University of Windsor for hosting the conference Web site. Finally, we thank the Organizing Committee and the members of the CSCSI Executive Committee for all their efforts in making AI 2007 a successful conference. May 2007
Ziad Kobti Dan Wu
Organization
AI 2007 was organized by the Canadian Society for the Computational Study of Intelligence (CSCSI).
Executive Committee General Chair Program Co-chairs
Tal Arbel (McGill University) Ziad Kobti and Dan Wu (University of Windsor)
Program Committee Referees Raza Abidi (Dalhousie University) Imran Ahmad (University of Windsor) Esma A¨ımeur (University of Montreal) Caroline Barri`ere (NRC) Shai Ben-David (University of Waterloo) Sabine Bergler (Concordia University) Cory Butz (University of Regina) Giuseppe Carenini (UBC) Yllias Chali (University of Lethbridge) David Chiu (University of Guelph) Douglas Dankel (University of Florida) Atilla El¸ci (E. Mediterranean University) Michael Fleming (University of New Brunswick) Richard Frost (University of Windsor) Scott Goodwin (University of Windsor) Robin Gras (University of Windsor) Jim Greer (University of Saskatchewan) Adlane Habed (University of Windsor) Malcolm Heywood (Dalhousie University) Robert Hilderman (University of Regina) Graeme Hirst (University of Toronto) Diana Inkpen (University of Ottawa)
Nafaa Jabeur (University of Windsor) Nathalie Japkowicz (Monash University) Grigoris Karakoulas (University of Toronto) Kamran Karimi (University of Lakehead) Vlado Keselj (Dalhousie University) Iluju Kiringa (University of Ottawa) Ziad Kobti (University of Windsor) Greg Kondrak (University of Alberta) Leila Kosseim (Concordia University) Luc Lamontagne (Laval University) Philippe Langlais (University of Montreal) Guy Lapalme (University of Montreal) Kate Larson (University of Waterloo) Fran¸cois Laviolette (Laval University) Pawan Lingras (Saint Mary’s University) Alejandro Lopez-Ortiz (University of Waterloo) Choh Man Teng (IHMC) Jean-Marc Mercantini (LSIS) Joel Martin (NRC) Bob Mercer (University of Western Ontario)
VIII
Organization
David Nadeau (NRC) Eric Neufeld (University of Saskatchewan) Roger Nkambou (University of Montreal) David Poole (UBC) Jian-Yun Nie (University of Montreal) Gerald Penn (University of Toronto) Fred Popowich (Simon Fraser University) Robert Reynolds (Wayne State University) Massih Reza Amini (Pierre and Marie Curie University) Luis Rueda (University of Concepcion) Anoop Sarkar (Simon Fraser University) Weiming Shen (NRC) Michel Simard (NRC) Stan Szpakowicz (University of Ottawa)
Ahmed Tawfik (University of Windsor) Nicole Tourigny (Laval University) Thomas Tran (University of Ottawa) Andre Trudel (Acadia University) Marcel Turcotte (University of Ottawa) Peter van Beek (University of Waterloo) Herna Viktor (University of Ottawa) Shaojun Wang (Wright State University) Dan Wu (University of Windsor) Yang Xiang (University of Guelph) Yiyu Yao (University of Regina) Jia You (University of Alberta) Li-Yan Yuan (University of Alberta) Harry Zhang (University of New Brunswick) Nur Zincir-Heywood (Dalhousie University)
Additional Reviewers Referees Tony Abou-Assaleh Dulce Aguilar-Solis Muath Alzghool Xiangdong An Alina Andreevskaia Jing Bai Susan Bartlett Yaohua Chen William Elazmeh Mark Eramian Oana Frunza S´ebastien Gambs Liqiang Geng Kevin Grant
Baohua Gu Franklin Hanshar Michael Horsch Zina Ibrahim Michael Janzen Sittichai Jiampojamarn Mehdi M. Kashani Abolfazl Keighobadi Lamjiri Alistair Kennedy Guohua Liu Haibin Liu Mehrdad Oveisi-Fordoei Robert Price
Behnam Rahnama Maxim Roy Elhadi Shakshuki Danny Silver Zhongmin Shi Jiang Su Hathai Tanta-ngai Jeff Taylor Javier Thaine Davide Turcato Bin Wang Hong Yao Haiyi Zhang
Sponsoring Institutions The Canadian Society for the Computational Study of Intelligence (CSCSI) ´ Soci´et´e Canadienne pour L Etude de l Intelligence par Ordinateur Ontario Centres of Excellence (OCE)
Table of Contents
Session 1. Agents Modeling Role-Based Agent Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yu Zhang
1
Distributed Collaborative Filtering for Robust Recommendations Against Shilling Attacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ae-Ttie Ji, Cheol Yeon, Heung-Nam Kim, and Geun-Sik Jo
14
Competition and Coordination in Stochastic Games . . . . . . . . . . . . . . . . . . Andriy Burkov, Abdeslam Boularias, and Brahim Chaib-draa
26
Multiagent-Based Dynamic Deployment Planning in RTLS-Enabled Automotive Shipment Yard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jindae Kim, Changsoo Ok, Soundar R.T. Kumara, and Shang-Tae Yee
38
R-FRTDP: A Real-Time DP Algorithm with Tight Bounds for a Stochastic Resource Allocation Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . Camille Besse, Pierrick Plamondon, and Brahim Chaib-draa
50
A Reorganization Strategy to Build Fault-Tolerant Multi-Agent Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sehl Mellouli
61
Session 2. Bioinformatics A Multi-site Subcellular Localizer for Fungal Proteins . . . . . . . . . . . . . . . . Michel Nathan
73
Selecting Genotyping Oligo Probes Via Logical Analysis of Data . . . . . . . Kwangsoo Kim and Hong Seo Ryoo
86
Session 3. Classification Learning the Semantic Meaning of a Concept from the Web . . . . . . . . . . . Yang Yu and Yun Peng
98
On Combining Dissimilarity-Based Classifiers to Solve the Small Sample Size Problem for Appearance-Based Face Recognition . . . . . . . . . Sang-Woon Kim and Robert P.W. Duin
110
A Novel Approach for Automatic Palmprint Recognition . . . . . . . . . . . . . . Murat Ekinci and Murat Aykut
122
X
Table of Contents
ICS: An Interactive Classification System . . . . . . . . . . . . . . . . . . . . . . . . . . . Yan Zhao, Yiyu Yao, and Mingwu Yan
134
Fast Most Similar Neighbor Classifier for Mixed Data . . . . . . . . . . . . . . . . Selene Hern´ andez-Rodr´ıguez , J. Francisco Mart´ınez-Trinidad, and J. Ariel Carrasco-Ochoa
146
Performance Measures in Classification of Human Communications . . . . . Marina Sokolova and Guy Lapalme
159
Cost-Sensitive Decision Trees with Pre-pruning . . . . . . . . . . . . . . . . . . . . . . Jun Du, Zhihua Cai, and Charles X. Ling
171
Probability Based Metrics for Locally Weighted Naive Bayes . . . . . . . . . . Bin Wang and Harry Zhang
180
Recurrent Boosting for Classification of Natural and Synthetic Time-Series Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robert D. Vincent, Joelle Pineau, Philip de Guzman, and Massimo Avoli Pattern Classification in No-Limit Poker: A Head-Start Evolutionary Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Brien Beattie, Garrett Nicolai, David Gerhard, and Robert J. Hilderman
192
204
Session 4. Constraint Satisfaction Managing Conditional and Composite CSPs . . . . . . . . . . . . . . . . . . . . . . . . . Malek Mouhoub and Amrudee Sukpan
216
Multiagent Constraint Satisfaction with Multiply Sectioned Constraint Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yang Xiang and Wanling Zhang
228
Session 5. Data Mining A Clustering Algorithm Based on Adaptive Subcluster Merging . . . . . . . . Jiani Hu, Weihong Deng, and Jun Guo
241
Efficient Algorithms for Video Association Mining . . . . . . . . . . . . . . . . . . . . B. SivaSelvan and N.P. Gopalan
250
Distributed Data Mining in a Ubiquitous Healthcare Framework . . . . . . . Murlikrishna Viswanathan
261
Constructing a User Preference Ontology for Anti-spam Mail Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jongwan Kim, Dejing Dou, Haishan Liu, and Donghwi Kwak
272
Table of Contents
Question Answering Summarization of Multiple Biomedical Documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zhongmin Shi, Gabor Melli, Yang Wang, Yudong Liu, Baohua Gu, Mehdi M. Kashani, Anoop Sarkar, and Fred Popowich A Profit-Based Business Model for Evaluating Rule Interestingness . . . . . Yaohua Chen, Yan Zhao, and Yiyu Yao
XI
284
296
Session 6. Knowledge Representation and Reasoning Reasoning About Operations on Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bernhard Heinemann
308
Analytic Results on the Hodgkin-Huxley Neural Network: Spikes Annihilation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dragos Calitoiu, John B. Oommen, and Doron Nussbaum
320
Improving Importance Sampling by Adaptive Split-Rejection Control in Bayesian Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Changhe Yuan and Marek J. Druzdzel
332
Adding Local Constraints to Bayesian Networks . . . . . . . . . . . . . . . . . . . . . Mark Crowley, Brent Boerlage, and David Poole
344
On the Use of Possibilistic Bases for Local Computations in Product-Based Possibilistic Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Salem Benferhat and Salma Smaoui
356
Reasoning with Conditional Preferences Across Attributes . . . . . . . . . . . . Shaoju Chen, Scott Buffett, and Michael W. Fleming
369
Path Propagation for Inference in Bayesian Networks . . . . . . . . . . . . . . . . . Dan Wu and Liu He
381
Problem-Solving Knowledge Mining from Users’ Actions in an Intelligent Tutoring System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Roger Nkambou, Engelbert Mephu Nguifo, Olivier Couturier, and Philippe Fournier-Viger Incremental Neighborhood Graphs Construction for Multidimensional Databases Indexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hakim Hacid and Tetsuya Yoshida
393
405
Session 7. Learning Learning Network Topology from Simple Sensor Data . . . . . . . . . . . . . . . . Dimitri Marinakis, Philippe Gigu`ere, and Gregory Dudek
417
XII
Table of Contents
Reinforcement Learning in Nonstationary Environment Navigation Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Terran Lane, Martin Ridens, and Scott Stevens
429
On the Stability and Bias-Variance Analysis of Kernel Matrix Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Vijaya Saradhi and Harish Karnick
441
Session 8. Natural Language Query-Based Summarization of Customer Reviews . . . . . . . . . . . . . . . . . . . Olga Feiguina and Guy Lapalme
452
Multi-state Directed Acyclic Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael Wachter and Rolf Haenni
464
Fuzzy Clustering for Topic Analysis and Summarization of Document Collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ren´e Witte and Sabine Bergler
476
Creating a Fuzzy Believer to Model Human Newspaper Readers . . . . . . . Ralf Krestel, Ren´e Witte, and Sabine Bergler
489
Rethinking the Semantics of Complex Nominals . . . . . . . . . . . . . . . . . . . . . . Nabil Abdullah and Richard A. Frost
502
A Hybrid Approach to Improving Automatic Speech Recognition Via NLP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kimberly Voll
514
Session 9. Planning Planning in Multiagent Expedition with Collaborative Design Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yang Xiang and Frank Hanshar
526
Hierarchical Shortest Pathfinding Applied to Route-Planning for Wheelchair Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Suling Yang and Alan K. Mackworth
539
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
551
Modeling Role-Based Agent Team* Yu Zhang Computer Science Department, Trinity University, San Antonio, TX 78212 Tel.:(210)999-7399, Fax:(210)999-7477
[email protected] Abstract. The problem of ensuring agents work as an effective team in dynamic distributed environments still remains a challenging issue. In this paper we proposed a role-based team model. In our model, the role characterizes the responsibilities and provides logic patterns to achieve certain goals and cooperate with others. The agent is an autonomous execution unit and follows the logic patterns that the role provides. We also developed algorithms and mechanisms to evolve the plan of a role to the plan of an agent. Our role-based team model allows the split of roles (who define the plans) and agents (who execute the plans) in team plans, and dynamic role-agent assignment. It also achieves a certain level of plan reusability. We present two experiments which show plan reusability and its flexibility in supporting simultaneously plan invocation. Keywords: Agent teamwork, Role, Plan.
1 Introduction Teamwork is becoming increasingly important in many dynamic multi-agent systems [13]. Agents in a team need to form joint mental states which drive agents to act together as a team and form the interactions leading their individual actions to team efforts [5, 7]. To simulate teamwork, a teamwork language is demanded to explicitly express the mental states underlying teamwork. In our opinion, the effective design of a teamwork language requires two aspects to consider. First, it should be able to handle unexpected uncertainties occurred in complex and dynamic domains, such as dynamic changes in team’s goals, team members’ unexpected failures to fulfill their responsibilities, decision-making in dynamic environment, and dynamically backing up other team members. Second, considering the perspective of software engineering, the teamwork language would better allow specify teamwork knowledge conceptually for being reused, particularly, team plans are better specified in terms of abstract entities, instead of specific agents, so as to be reused by different teams of agents. A lower level of abstraction, role, is currently used by many researcher of multiagent systems to close this gap [14, 9, 11, 6, 16]. Biddle and Thomas’s role theory views role as the concept of partitioning behaviors and emphasizing coordination and cooperation [2]. Becht’s ROPE (Role Oriented Programming Environment for multiagent systems) uses roles to decouple the organization of agents from the structure of cooperation processes [1]. Cooperation process is designed from a global perspective *
This work was supported by DoD MURI F49620-00-I-326 administered through AFOSR.
Z. Kobti and D. Wu (Eds.): Canadian AI 2007, LNAI 4509, pp. 1–13, 2007. © Springer-Verlag Berlin Heidelberg 2007
2
Y. Zhang
and largely independent of concrete agents so that shifting cooperative behavior does not require changing agents (agents can fill multiple roles and switch between them). Stone and Veloso introduced roles as a mechanism for specifying an agent’s internal and external behaviors and decomposing team tasks [12]; they then used this to model robot soccer. A formation decomposes the task space by defining a set of roles. There are as many roles as there are agents in the team, so that each role is filled by one agent. The mapping between agents and role is not pre-specified. In this paper, we propose a role-based teamwork language. Different with the existing work, we use roles and role variables distinguish static (by roles) and dynamic (by role variables) action associations; and when delegating roles and role variables in a plan to agents, we have the agents form s joint mental state to enforce the execution of the plan as a team effort, particularly the sub-actions in the plan will be executed coherently. Our concepts of role and role variable enable our mechanisms of task decomposition and delegation, by which role-based plans drive agents to actually execute teamwork. Our mechanism of task decomposition is based on a notion of responsibility, which is defined in terms of what a responsibility contains and how a responsibility impacts the mental states of the agent(s) taking the responsibility. Our mechanism of task delegation has three steps: 1) a team task is translated to a team responsibility which is represented by a graph; 2) through decomposing the team responsibility graph to individual responsibility graphs, a team task is decomposed to individual sub-tasks; and 3) individual sub-tasks are delegated to agents. The structure of this paper is as follows. Section 2 formalizes the basic concepts. Section 3 introduces algorithms for task decomposition and delegation. Section 4 introduces delegating roles to agents, formalize a notion of admissible assignment, and present a CSP algorithm to search for admissible assignments. Section 5 is experiment. Section 6 concludes the paper.
2 Role-Based Plan In their role theory, Biddle and Thomas concluded that roles can be defined based on partitioning concepts for persons and their behaviors [2]. They also pointed out that pre-association with roles is too restrictive, i.e. the determination of which role actually performs which action being determined dynamically in a specific situation. Based on it, we define a position as all entities that can perform a set of primitive operations, and use this to characterize a collectively recognized category of persons who are able to exhibit a set of behaviors. We define a role as an entity of a position associated with a bag of temporally ordered actions. We define a role variable as an entity which is dynamically selected from a set of roles to be associated with a bag of temporally ordered actions. Generally speaking, a role variable stands for some role out of a set of roles. Depending on concrete situation, one role from the set is dynamically selected to fill the role variable. When this happens, the bag of operations of the assuming role is dynamically expanded. 2.1 Position We define a concept of position to refer a collectively recognized category of entities (persons) that exhibit a set of behaviors.
Modeling Role-Based Agent Team
3
Definition 1. An entity is an abstraction of any performer (e.g., agent or person) that is able to perform operations. Definition 2. A position is a named set of all entities that are able to perform a set of primitive operations, denoted by operators. Given a set of operations O, a position based on the operations is P(O)={e|e∈ Entities ∧ ∀o (o∈O) ∧ Capable(e, o)},
(1)
where e represents an entity and Capable(e, o) means that e is able to perform operation o, assuming the preconditions of o are true. The purpose of defining position is to capture the capability requirements on agents. Every entity of a position is required to be capable to do the operators defined in the position. For example, (POSITION sniffer (talk move movein randmove sense selectTarget)).
(2)
A position sniffer is defined by a set of operators, including talk, move, movein, randmove, sense, and selectTarget. Suppose e is an entity that takes on the position sniffer; e should be able to perform the above operators. 2.2 Role We base our definition for a role on a position. In other words, any action (i.e., primitive operation) associated with the role must be in the operation set of the position. Let O be a bag of operations that must be performed by the same entity of a position, COND be a bag of conditional operators where each action is contingent on a conjunction of conditions, and CO be a bag of ordering constraints that impose a “temporal order” on O and the evaluation of the conditions in COND. We define a role r as an abstraction of the entity of position P that satisfies the constraints. Definition 3. A role is an abstraction of an entity that performs a specific bag of operations and includes temporal constraints on the order in which the operations may be performed: r = (id, P, O, COND,
