New Frontiers in Artificial Intelligence: JSAI 2006 Conference andWorkshops (Lecture Notes in Computer Science, 4384)

✍ Scribed by Takashi Washio (editor), Ken Satoh (editor), Hideaki Takeda (editor), Akihiro Inokuchi (editor)

• Publisher: Springer
• Year: 2007
• Tongue: English
• Leaves: 401
• Category: Library

✦ Synopsis

The progress in information technology including arti?cial intelligence (AI) in the last few decades is remarkable, and has attracted many young researchers to this ?eld. This trend is now accelerated along with the recent rapid growth of computer communicationnetworksandthe worldwideincreaseinresearchers. In this context, we have observed many outstanding AI studies in Japanese dom- tic conferences. They have high technical originality, quality and signi?cance. The annual conference of JSAI (Japan Society for Arti?cial Intelligence) is one of the key and representative domestic meetings in the ?eld of intelligent inf- mation technology. Award papers in this conference have an excellent quality of internationalstandards. The annualconferenceofJSAIalsoorganizesco-located international workshops to provide excellent study reports. The objectives of this book are to present the award papers of the 20th annual conference of JSAI 2006 and selected papers from the three co-located international workshops and to promote the study exchange among researchers worldwide. Eight papers were awarded among more than 200 presentations in the conference, and 21 papers were selected from a total of 44 presentations in the workshops of Logic and Engineering of Natural Language Semantics 2006 (LENLS 2006), Learning with Logics and Logics for Learning (LLLL 2006) and Risk Mining (RM 2006). The award papers in the 20th annual conference of JSAI 2006 were selected from presentations covering the wide ?eld of art- cial intelligence through the processes of candidate recommendations, detailed open discussions and voting by ProgramCommittee members of the conference.

✦ Table of Contents

Cover
Title Pages
Preface
Organization and Editorial Board
Table of Contents
Overview of Awarded Papers – The 20th Annual Conference of JSAI
Translational Symmetry in Subsequence Time-Series Clustering
Introduction
Lattice Model for Time Series Analysis
Vector Space in Dirac's Notation
Linear Operators in ${\cal H}_0$
Lattice Model for Time Series Data
Translation Operator on the Lattice
Spectral Clustering of Subsequences
Group-Theoretical Properties of $ \hat{ \rho }$
Translational Group
Representation Theory
Irreducible Representations of Translational Group
Solution to Subsequence Time-Series Clustering
The $w = n$ Case
The $w < n $ Case
Concluding Remarks
Visualization of Contents Archive by Contour Map Representation
Introduction
Knowledge Landscape
Sustainable Knowledge Globe
Contour Map Representation
Nesting Contour
Dendroidal Contour
Island-Like Contour
Discussion
Conclusion
Discussion Ontology: Knowledge Discovery from Human Activities in Meetings
Introduction
Discussion Ontology
Gathering Metadata for Generating Discussion Content
Structuring Discussions
Discussion Mining
Usage of Discussion Content
Classification of Statements and Extraction of Discussion Flow
Intention Tag
Classification of Statements
Extraction of Discussion Flow Based on Statement Intention
Conclusion and Future Work
Predicting Types of Protein-Protein Interactions Using a Multiple-Instance Learning Model
Introduction
A Model of Interaction Between a Protein Complex Pair
A Method for Predicting Interaction Types Based on Multiple Instance Learning
Multiple-Instance Learning Scheme
Diverse Density
Modified Diverse Density as a Weighted Voting System
PPI Dataset and Feature Space
PPI Dataset Obtained from the KEGG Pathways
Feature Space Encoding the Gene Ontology
Experiments: Binary Classification Task Concerning Phosphorylation
Binary Classification Task Concerning Phosphorylation
Results
Conclusion
References
Lattice for Musical Structure and Its Arithmetics
Introduction
Relative Pseudo-complement and Heyting Calculus
Algebra and Lattice
Relative Pseudo-complement
Heyting Calculus
Calculation of Relative Pseudo-complement
Efficient Algorithm
Value of Relative Pseudo-complement
Abstraction of Distance Between Musical Fragments
Definitions of Meet and Join Operations
Applicable Domain of Direct Proportion Algorithm
Formal Representation of Media Arrangement
Feature Structure as Media Representation
Representation of Time-Span Tree
Arrangement by Tensor Product
Concluding Remarks
Viewlon: Visualizing Information on Semantic Sensor Network
Introduction
Semantic Sensor Network
Classes
Instances
Sensors
Inference Rules
Related Work
Visuali Zing Sensor Data
Visualizing Information Related to Ontology
Viewlon
System Design
Views and Panels Implemented in Viewlon
Information Display Levels
Timeline View
Processing Queries
Experiment
Experiment of Visualizing Functions
Experiment on Processing Queries
Convergence Time of Nodes in Main View
Conclusion
Cooperative Task Achievement System Between Humans and Robots Based on Stochastic Memory Model of Spatial Environment
Introduction
Stochastic Memory Model of Spatial Environment
Spatial Memory Model for Environments and Objects
Revision of Spatial Memory by Vision
Reflection of Context for the Memory Model
Application for Behavior Decision and Dialogue Control
Cooperative Behavior Using Probabilistic Geometric Memory Model
An Experiment of Behavior Control and Understanding of Context by Memory
Representation of Usual Context in Daily-Life Environment
Conclusion
People Who Create Knowledge Sharing Communities
Web-Based Knowledge Sharing Community
Yahoo! Chiebukuro
Methods
Procedure
Questionnaire
Objective Behavioral Index
Results
Demographic Traits
Participation in the Community
Reason for Question-Posting
Reason for Answer-Posting
Discussion
Conclusion
Logic and Engineering of Natural Language Semantics (LENLS) 3
The Workshop
Papers
A Dynamic Semantics of Intentional Identity
Introduction
On Intentional Identity and Counterpart Relation
Geach's Hob-Nob Sentence
Edelberg's Arsky-Barsky Sentences
Cross Speaker Anaphora
Weak Intentional Identity
Multiple Intentional Identity
Edelberg's Semantics
A Dynamic Semantics
Conclusion
Prolegomena to General-Imaging-Based Probabilistic Dynamic Epistemic Logic
Introduction
Probabilistic Epistemic Logic $\mathsf{PEL}$
Language
Semantics
Syntax
Soundness and Completeness
Conditionalisation-Based Probabilistic Dynamic Epistemic Logic $\mathsf{CPDEL}$
Language
Semantics
Syntax
Monty Hall Dilemma
Semantic Analysis of Monty Hall Dilemma in Terms of $\mathsf{CPDEL}$
General-Imaging-Based Probabilistic Dynamic Epistemic Logic $\mathsf{GIPDEL}$
Language
Semantics
Syntax
Soundness and Completeness
Semantic Analysis of Monty Hall Dilemma in Terms of $\mathsf{GIPDEL}$
Semantic Analyses of Modified Version of Monty Hall Dilemma
Modified Version of Monty Hall Dilemma
Semantic Analysis in Terms of $\mathsf{CPDEL}$
Semantic Analysis in Terms of GIPDEL
Concluding Remarks
References
Logical Dynamics of Commands and Obligations
Introduction
The Static Base Logic $\mathsf{MDL^{+} II}$
A Dynamic Extension $\mathsf{ECL \, II}$
Proof System for $\mathsf{ECL \, II}$
Built-In Assumptions and Interesting Validities and Non-validities
Conclusion
On Factive Islands: Pragmatic Anomaly vs. Pragmatic Infelicity
Introduction
Basic Facts
Types of Wh-Phrases That Are Subject to the FIC
Unique wh-Phrase
Previous Analyses and Their Problems
Syntax-Based Approach
Semantics-Based Approach
An Alternative Analysis
Some Ontological Assumptions and Terminological Clarifications
Unique wh-Phrase
WHY
HOW
Pragmatic Anomaly vs. Pragmatic Infelicity
Aspects of the Indefiniteness Effect
Introduction
The Syntactic Aspect
The Semantic Aspect
The Pragmatic Aspect
Concluding Remarks
Interpreting Metaphors in a New Semantic Theory of Concept
Motivation
The Intensionality of Metaphors
The Nonreversibility of Metaphors
The Partial Systematicity of Metaphors
The Context-Sensitivity of Metaphors
The Coexistence of Literal and Nonliteral Meanings
The Death of Metaphors
Semantic Interpretation of Concept
Concept and Metaphor
A Formalism of Metaphors
Syntax
Semantics
Conclusions
Covert Emotive Modality Is a Monster
Shifted Context in Unembedded Sentences
Fake Past and Emotive Modality
Data
Fake Past
“Fake” Operator Shifts Context Parameters
Negative Presupposition
Bouletic and Speculative Modality
The Phrase Structure of Fake Past
Narrative Fake Present
Data
Context Change by Narrator’s Report Function
Discourse Initial \emph{Also/Too}
The Third \emph{Mo} also/even' Negative and Positive Emotions Korean \emph{To}also/even' and Chinese \emph{Ye} `also'
Sentence Focus and Presuppositions
Bouletic Modality
Information Update
Conclusion
References
Conversational Implicatures Via General Pragmatic Pressures
Introduction
A System of Pragmatic Pressures
Probabilities and Propositions
Quality Thresholds
Quantity Ratings
Relevance Ranking
Felicitous Utterances
Conversational Implicatures
Relevance-Based Implicatures
Quantity-Based Implicatures of Questions
Quantity-Based Implicatures of Answers
Summary and Prospects
Dake-wa: Exhaustifying Assertions
Introduction
Contrastive-Marking
Puzzle
Levels of Meaning Contributed by $dake$
Scope of $Dake$
Exhaustification over Propositions
Quantification at Non-propositional Levels
Exhaustification over Potential Literal Acts
Conclusion
Unembedded ‘Negative’ Quantifiers
Introduction: Unembedded Quantifiers
Setups
$KANENAS$ in Greek and $ni-kogo$ in Russian
Context
Anaphoricity
Unembedded $KANENAS$ and $ni-kogo$
Unembedded $ni-kogo$
Unembedded $KANENAS$
Conclusion and Remaining Issues
References
The Fourth Workshop on Learning with Logics and Logics for Learning (LLLL2006)
The Workshop
The Post-workshop Proceedings
Consistency Conditions for Inductive Inference of Recursive Functions
Introduction
Preliminaries
Coherence and Consistency of Learning Strategies
Characterizations
Hierarchy Results
Conclusions and Future Work
Inferability of Closed Set Systems from Positive Data
Introduction
Inferability from Positive Data
Inferability of Closed Set Systems
Inferability of Unions of Closed Set Systems
Inferability of Algebraic Closed Set Systems
Discussion and Conclusion
An Extended Branch and Bound Search Algorithm for Finding Top-N Formal Concepts of Documents
Introduction
Preliminaries
Finding Document Clusters Based on Formal Concept Analysis
Document Clusters as Formal Concepts
Algorithm for Finding Top-$N$ $\delta$-Valid $FC$-Clusters
Experimental Result
Discussion
Concluding Remarks
N-Gram Analysis Based on Zero-Suppressed BDDs
Introduction
ZBDDs and Item Set Manipulation
Sets of Combinations and Their Representation
Item Set Histograms and ZBDD Vectors
ZBDD-Based Representation for Sets of Sequences
Sets of Sequences and Sets of Combinations
$n$-Gram Representation
Binary Coding of Symbols
Experimental Results
Generating $n$-Gram Data Based on ZBDDs
Post Processing for $n$-Gram Data
Conclusions
Risk Mining - Overview
Introduction
Workshop Overview
Program Committee
Analysis on a Relation Between Enterprise Profit and Financial State by Using Data Mining Techniques
Introduction
Objective Data
Model Selection for Profit Prediction
Outline of Random Tree Based Committee Learning
Random Tree
Committee Learning
Characteristics of Financial Settlement Data
Discussion and Conclusion
Unusual Condition Detection of Bearing Vibration in Hydroelectric Power Plants for Risk Management
Introduction
Measurement Data
Abnormal Condition Sign Detection
One Class SVM
Unusual Pattern Detection Experiment
Experimental Setup
Detection Results and Evaluation
Inexplainable Unusual Condition Patterns
Conclusion
Structural Health Assessing by Interactive Data Mining Approach in Nuclear Power Plant
Introduction
Conceptual Framework of Interactive Data Mining Method
Adverseness of Neural Network in Interpreting Plant Quake-Proof Data
Interactive Data Mining Approach
Architecture of Information Inference System
Application: Detection of Structural Damage
Problem Identification and Data Generation
Detail of Data Classifier
Training and Test Result Analysis
Concluding Remarks
References
Developing Mining-Grid Centric e-Finance Portals for Risk Management
Introduction
The Proposed Approach
Portal Architecture
Data-Grid vs. Data Warehouse
Mining-Grid vs. OLAP
Knowledge-Grid for Knowledge Management
Security Concerns
Security Objective
Three-Level Security
A Case Study: Credit Card Risk Management
POM Agent
MVF Agent
Learning Active Knowledge on the Knowledge-Grid
Related Work
Conclusions
Knowledge Discovery from Click Stream Data and Effective Site Management
Introduction
Explanation of Data and Data Pre-processing
Basic Analysis of All the Shops in the Projects
Purchase Behavior on the Sites
Purchase Probability and Number of Sessions
A Site That Can Bring in Wavering Potential Customers and Make Them Buy
Defining the Target Customers
Extracting the Features of the Decision Tree
Tracking Analysis Using the Character String Analysis System, E-BONSAI
Implications for Business
Conclusions
Sampling-Based Stream Mining for Network Risk Management
Introduction
Related Work
Requirements on Stream Mining
Sampling-Based Stream Mining
Experimental Results
Discussion
Conclusion
Relation Between Abductive and Inductive Types of Nursing Risk Management
Introduction
Inductive Nursing Risk Management
Necessity of Risk Management and Data Mining in Hospitals
Features of Accident and Incident Report Databases
Data Collection Through Experiments in Hospitals
Abductive Nursing Risk Management
Abduction Model
Scenario Violation Model
Relation Between Abductive and Inductive Types of Risk Management
Features of Both Types of Nursing Risk Management
Cooperation Between Both Types of Nursing Risk Management
Conclusions
Author Index