Download AI*IA 2001: Advances in Artificial Intelligence: 7th Congress of the Italian Association for Artificial Intelligence Bari, Italy, September 25–28, 2001 Proceedings PDF

TitleAI*IA 2001: Advances in Artificial Intelligence: 7th Congress of the Italian Association for Artificial Intelligence Bari, Italy, September 25–28, 2001 Proceedings
File Size7.2 MB
Total Pages407
Table of Contents
                            Front matter
	AI*IA 2001: Advances in Artificial Intelligence
	Organizing Committee
	Table of Contents
Chapter 1
	Hard Relational Problems
	A Stochastic Approach
	Evaluation the Hypothesis Space Size
		Estimating the Hypothesis Space Size for a Function Free Language
		Estimating the Frequency of Concept Generalizations
	An Experimental Evaluation
Chapter 2
	1  Introduction
	2  Monte Carlo Algorithms
	2.1  Biased Monte Carlo Algorithms
	3  Relations between Majority Voting and Monte Carlo Algorithms
	4  Experimental Setting
	4.1  Weak Learner
	4.2  Datasets
	4.3  Experimental Results
	5  Conclusions
Chapter 3
	1	Introduction
	2. Background and Motivation
	3. Stepwise Construction of Model Trees
	4. Observations on Experimental Results
	4.1 Effect of Node Weighting
	4.2 Comparison with Other Systems
	4.3 Experiments on Laboratory-Sized Data Sets
	5. Conclusions
Chapter 4
	The Probabilistic Model for Hypertexts
	Evaluation Methods for Hyperlinks
	Experimental Results
Chapter 5
	The ANN Architecture
	The ANN-FR Learning Algorithm
	Application Examples
		Iris Dataset
		Heart Disease Dataset
Chapter 6
	Experimental Data Sets
		MR Images
		Remote Sensed Data
	On the Selection of Basis Function Parameters
	Clustering Strategy
Chapter 7
	Support Vector Machines
		Multi-class Classification with Error Correcting Codes
	Recursive Neural Networks
	Experimental Results
		Stacked Integration of Flat and Structural Classifiers
		Results with SVMs
		Combining Flat and Structural Features with SVM and ECC
Chapter 8
		Previous Works on Automatic Intelligent Accompanist
	The Intelligent Accompanist
		Modeling the Signal
		Modeling the Performance
		Training the HMM
		Playing the Accompaniment
	Experimental Results
	Conclusions and Future Work
Chapter 9
	The Weighted Weak Learner
	Experimental Setting
	Experimental Results
	Discussion and Conclusions
Chapter 10
	The Learning System
	Handling Text in Document Processing
	Conclusions & Future Work
Chapter 11
	1.	Introduction
	2.	User Profiling
	3.	Profile Extractor Module Architecture
	4.	Profile Extractor Submodules
	4.1	Learning Problem Manager Submodule
	4.2	Profile Rule Extractor Submodule
	4.3	Classifier Submodule
	5.	Experimental Results
	5.1	First Experiment
	5.2	Second Experiment
	6.	Conclusions
Chapter 12
	A General Framework for Mereotopology
	Computing Similarities
	The {tt IRIS} Prototype
	Concluding Remarks
Chapter 13
		textsc {QuBE}{}
		Testing Methodology
	Backjumping vs. Trivial Truth
	textsc {QuBE}{} and State-of-the-Art QBF Solvers
Chapter 14
	Preliminaries on Logic Programming
		Stable Model Semantics
	A Model of Abduction with Penalization
	An Example: The Traveling Salesman Problem
	Computational Complexity
		Preliminaries on Complexity Theory
		Complexity Results
Chapter 15
	Qualitative Deviations
	Causal Reasoning
		Causal Semantics
		Using Causality to Improve Reasoning
	The Example System
		The Common Rail System
		The Qualitative Deviations Causal Model
	Discussion and Related Work
Chapter 16
	Local Search and Parallelism
	Parallel Local Search for MAX-SAT
	Experimental Results
	Conclusion and Future Work
Chapter 17
	Chemical Formulation for Product Revise
	The Design of Rubber Compounds
		Rubber Compounds for Motorsport Car Tires
		Applying the CBR Approach
	Concluding Remarks
Chapter 18
	Belief Revision
Chapter 19
	[email protected] {FPC}$: A Conceptual Modelling Language
	System Functions and Problem Solving Capabilities
	A Configuration Example
Chapter 20
	Encompassing Semantics in the Conceptual Model
		Integrating Diverse Ontologies
		Reasoning with Ontologies
		Nature of Ontologies
	A Knowledge Model Representing the Enriched Conceptual Model
	Relating the Extended Knowledge Model to the Motivations
		Behaviour over Time
		Prototypes and Exceptions
	A Modelling Example
Chapter 21
	Overview of Agent Architecture
		Camera Agent
		Object Agent
		Behavioural Classification
	Scene Modelling Using HMMs
		Markov Models for Scene Understanding
		Behavioural Classification
	Answering the Reviewers' Comments
Chapter 22
	1 Introduction: Agents' Autonomy
	2 Delegation/Adoption Theory
	3 The Adjustment of Delegation/Adoption
	4 The Adjustment of Delegation/Adoption
	5 Principles and Reasons for Autonomy Adjustment
	6 Conclusions
	7 References
Chapter 23
	Agents, Actions, and Obligations
	Legal Relations
	A Formal Definition of Obligations
	Directed Obligations
	Obligations and Legal Relations
Chapter 24
	The Logic
		A Product Version of $DLTL$
	Action Theories
Chapter 25
	emph {L*MASS}
		The Territorial MASS
		The Strategic MASS
Chapter 26
	What Is a `Deceptive' Communication?
	The Art of Deception
	A Deception Simulator: Mouth of Truth
	Which Beliefs to Ascribe to A
		Creating a Structure for IM-S-A
		Assigning Parameters to IM-S-A
	How to Select a Communicative Act
		Listing Candidates
		Analysing the Impact of Candidates on the Deception Object
		Analysing the Plausibility of Candidates
		Supporting the Communication with a `Reliable' Information Source
	An Example: The Tale of ``Scratchy and Itchy"
	Concluding Remarks
Chapter 27
	The Virtual Market Place
	Experimental Results
Chapter 28
	1 	Introduction
	3	Linguistic Expansions
		3.1 Semantic Disambiguation
		3.2 Answer Type Identification
		3.3 Keywords Expansion
Search Component
Conclusion and Future Work
Chapter 29
	1   Introduction
	2 Understanding the Image
	3 Generation of Image Descriptions
	4	User Modeling
	5  Planning the Image Description
	6	Rendering the Image Description
	7	Comparing Images
	8 	Conclusions and Future Work
Chapter 30
	Incremental Parsing with Connection Paths
		Parsing Strategy
	Formulating Parse Decisions with a Neural Network
		The Learning Task
		Neural Network Architecture
	Implementation and Results
		Results on Learning First Pass Attachments
		Preliminary Results on Incremental Parsing
Chapter 31
	A Flexible Framework for Parser Design
		Extended Dependency Graph
		Composition of Parsing Modules
		Lexicalised Syntactic Processors
	Evaluating Alternative Parsing Architectures
Chapter 32
	1	Introduction
	2 	‘User in Context’ Models
	3 	Presentation Aspects to Be Varied
	4	Our Presentation Agent
	5	An Example
	5.1.	Procedural Information Presentation
	6	Conclusions
Chapter 33
	Natural Language Processing and Text Classification
	A Hybrid Feature Selection Model
		The Problem of Feature Selection
		Generalizing Rocchio Formula for Selecting Features
		The Role of NLP in Feature Extraction
	Experimenting Hybrid Feature Selection in Text Classification
Chapter 34
	Knowledge Representation and Anchoring
		The Anchoring Process
	Extension to Multi-agent Systems
Chapter 35
	Our Model of Visual Perception: An Overall View
	Conceptual Spaces for Representing Motion
	Mapping Symbolic Representations on Conceptual Spaces
	Some Conclusions
Chapter 36
	HOPS: A Hybrid Omnidirectional/Pin-Hole Sensor
	Obstacle Detection with a Hybrid Vision System
		Computing Stereo Disparity
		Removing Perspective with HOPS
		Obstacle Detection with HOPS
	Real-World Experiments
	Final Remarks and Future Applications of HOPS
Chapter 37
	Map-Based Model Knowledge
	Modelling System
		Calibration and Estimation of the Projection Matrix
		Using Map Constraints
Chapter 38
	Theoretical Remarks
	Description of the System
	Experimental Results
Chapter 39
	A CSP-based Software Architecture
	The Constraint Data-Base
		Multi-capacity Resources
	Constraint-Guided Problem Solving
	The Plan Execution Monitor
	An Interaction Module for Continuous Support
		Current Visual Tools
	Concluding Remarks
Chapter 40
	Design of a Planning Architecture
		The Constraint Solver
		The Information Gathering Module
		The Generative Planner
		Reactive Executor
Chapter 41
	A Layered Architecture for a Computer Game
	Planning and Abstraction Hierarchies
	Vertical Interactions
	Horizontal Interactions
	Conclusions and Future Work
Back matter
	Author Index
Document Text Contents
Page 1

Lecture Notes in Artificial Intelligence 2175
Subseries of Lecture Notes in Computer Science
Edited by J. G. Carbonell and J. Siekmann

Lecture Notes in Computer Science
Edited by G. Goos, J. Hartmanis, and J. van Leeuwen

Page 2

New York
Hong Kong

Page 203

192 Valentina A.M. Tamma and Trevor J.M. Bench-Capon

viewpoint an ontology is an a priori description of what constitues necessary
truth in any possible world [Kri80]. Such a formal standing on ontologies per-
mits to add a meta-level of description to ontologies and thus to reason about
meta-properties [GW00]. We believe that in order to be able to share and reuse
ontologies and to reason with the knowledge expressed in ontologies, the for-
mal meta-level of the description should be complemented by a richer concept
description, more oriented to the knowledge sharing task. If we consider the dif-
ferent ways in which the term ontology has been used in artificial intelligence, we
obtain a spectrum where formal ontologies are at one end, while something close
to knowledge bases are at the other end of the spectrum. Our view on ontol-
ogy is somewhere in the middle: ontologies should provide enough information
to enable knowledge engineers to have a full understanding of a concept as it
is in the actual world, but should also enable knowledge engineers to perform a
formal ontological analysis. For this reason, we believe in ontologies that provide
an a priori account of necessary truth on all the possible worlds but also some
information on the actual world and all the worlds accessible from it.
A full understanding of a concept involves more than this, however: it is impor-
tant to recognise which properties are prototypical [Ros75] for the class mem-
bership and, more importantly, which are the permitted exceptions. There are,
however differences in how confident we can be that an arbitrary member of
a class conforms to the prototype: it is a very rare mammal that lays eggs,
whereas many types of well known birds do not fly. Understanding a concept
also involves understanding how and which properties change over time. This
dynamic behaviour also forms part of the domain conceptualisation and can
help to identify the meta-properties holding for the concept.

3 A Knowledge Model Representing the Enriched
Conceptual Model

In this section we illustrate a frame-based model which results by representing
the elements of the conceptual model in terms of the frame paradigm. We have
chosen to extend a frame-based, OKBC-like [CFF+98] knowledge model, since
the frame-based paradigm applied to ontologies is thought of being easy to use
because closer to the human way of conceptualise, and providing a rich expressive
power (a discussion on frame-based languages for ontologies can be found in
In this model properties are characterised with respect to their behaviour in the
concept description. The knowledge model is based on classes, slots, and facets.
Classes correspond to concepts and are collections of objects sharing the same
properties, hierarchically organised into a multiple inheritance hierarchy, linked
by IS-A links. Classes are described in terms of slots, or attributes, that can
either be sets or single values. A slot is described by a name, a domain, a value
type and by a set of additional constraints, here called facets. Facets can contain
the documentation for a slot, constrain the value type or the cardinality of a
slot, and provide further information concerning the slot and the way in which

Page 204

Characterising Concept’s Properties in Ontologies 193

the slot is to be inherited by the subclasses. The set of facets provided by OKBC
has been extended in order to encompass descriptions of the attribute and its
behaviour in the concept description and changes over time. The facets we use
are listed below, where we distinguish epistemic nature facets from ontological
nature ones, and are discussed in the next section:

– Defining Values: It associates a value v ∈ Domain with an attribute in
order to represent a property. However, when the concept that is defined is
very high in the hierarchy (so high that any conclusion as to the attribute’s
value is not possible), then it is more likely to associate with the slot Defin-
ing Values either the whole domain (when no decision at all can be made
on the attribute’s value) or a subset of the domain (when a concept is defined
by means of inheritance from a parent, thus the concept inherits the slot’s
filler from its parent but specialises it by identifying a subset of the domain
characterising the parent concept), that is Defining Values = Domain or
Defining Values = Subdomain ⊂ Domain.For example, when describing a
generic concept such as Person in terms of the attribute Age, we can asso-
ciate with this slot the Defining Values=[0, 120], expressing the fact that
a person’s age can range between 0 and 120. In such a case [0, 120] coincides
with Domain. The concept is too generic in order to associate a single value
with the slot Age. If, then, we define the concept Teenager as subconcept
of Person, this inherits from Person the slot Age, but the child concept is
qualified by associating with this slot the Defining Values=[11, 18] which
is a subset of [0, 120]. This is an ontological facet;

– Value descriptor: The possible filler for this facet are Prototypical, Inher-
ited, Distinguishing, Value. An attribute’s value is a Prototypical one if the
value is true for any prototypical instance of the concept, but exceptions
are permitted with a degree of credibility expressed by the facet Modality.
An attribute’s value can be Inherited from some super concept or it can be
a Distinguishing value, that is a value that differentiates among siblings. If
this facet is set to Value this means that the value is neither prototypical,
nor inherited or distinguishing. Note that inherited and distinguishing values
are incompatible in the same concept description, that is a value is either
inherited or distinguishing, but cannot be both. On the other hand a value
can be prototypical and inherited. Distinguishing values become inherited
for subclasses of the class. This is an ontological facet;

– Exceptions: It can be either a single value or a subset of the domain. It indi-
cates those values that are permitted in the concept description because they
are in the domain, but deemed exceptional from a common sense viewpoint.
The exceptional values are not only those which differ from the prototypical
ones but also any value which is possible but highly unlikely. This property
is epistemic;

– Modality: An integer describing the degree of confidence of the fact that the
attribute takes the value specified in the facet Value. It describe the class
membership condition. The possible values are 1: All, 2: Almost all, 3: Most,
4: Possible, 5: A Few, 6: Almost none, 7: None. The value None associated

Page 406

Author Index

Abbattista, F., 87
Adorni, G., 344
Aiello, M., 99
Andreoli, C., 285
Appice, A., 20
Armano, G., 388

Bandini, S., 159, 249
Baraldi, A., 51
Barruffi, R., 382
Basili, R., 308, 320
Bellino, A., 20
Bench-Capon, T.J.M., 189
Blonda, P., 51
Blum, C., 147
Boella, G., 225
Bolognini, L., 344
Bonarini, A., 327
Botta, M., 70

Cagnoni, S., 344
Carofiglio, V., 255
Cassotta, M.T., 285
Castelfranchi, C., 212, 255
Castellano, G., 40
Ceci, M., 20
Cesta, A., 369
Chella, A., 333, 356, 362
Cherchi, G., 388
Cortellessa, G., 369
Costa, F., 297

D’Addabbo, A., 51
De Blasi, R., 51
De Carolis, B., 285, 314
De Cicco, M.L., 285
Di Maggio, P., 314

Esposito, R., 11

Falcone, R., 212
Fanelli, A.M., 40
Fanizzi, N., 81, 87
Favali, L., 225
Ferilli, S., 81, 87

Frasconi, P., 33, 57, 297
Frixione, M., 333

Gaglio, S., 333
Giordana, A., 1
Giordano, L., 165, 237
Giunchiglia, E., 111
Gliozzi, V., 165
Grassano, R., 255
Guarino, M.D., 362

Infantino, I., 356

Jones, G.A., 201

Leone, N., 123
Lesmo, L., 225
Lombardo, V., 297
Lops, P., 87

Magnini, B., 273
Magro, D., 176
Malerba, D., 20
Manzoni, S., 159, 249
Marcialis, G.L., 57
Martelli, A., 237
Matteucci, M., 327
Milano, M., 382
Monekosso, N., 201
Mordonini, M., 344
Moschitti, A., 320

Narizzano, M., 111
Negri, M., 273
Neri, F., 267

Oddi, A., 369
Olivetti, N., 165
Orio, N., 64

Pallotta, D., 20
Panati, A., 135
Passerini, A., 33
Pavesi, G., 249
Pazienza, M.T., 308, 320
Perri, S., 123
Pirrone, R., 362

Page 407

396 Author Index

Pizzutilo, S., 314
Policella, N., 369
Pontil, M., 57
Prevete, R., 273

Remagnino, P., 201
Restelli, M., 327
Roli, A., 147
Roli, F., 57
de Rosis, F., 255, 285

Saitta, L., 11
Scarcello, F., 123
Schwind, C., 237
Semeraro, G., 81, 87
Serra, A., 1

Simone, C., 249
Soda, G., 33, 297
Susi, A., 369

Tacchella, A., 111
Tamma, V.A.M., 189
Tarantino, C., 51
Theseider Dupré, D., 135
Torasso, P., 176

Vargiu, E., 388

Yao, Y., 57

Zanzotto, F.M., 308

Similer Documents