Inference by Local Computation- CANCELLED -

Qualitative Spatial and Temporal Reasoning: from theory to applications- CANCELLED -

Argumentation and Dialogue in AI

Specification and verification of epistemic properties in multi-agent systems- CANCELLED -

Constraint Processing

Stochastic Local Search: Foundations and Applications- CANCELLED -

DataMining: From rule learning to relational subgroup discovery

Information Extraction: Theory and Practice- CANCELLED -

Semantic Web: Vision, Tools and Applications

T01

Inference by Local Computation.

CANCELLED

Jürg Kohlas, Jérôme Mengin, Nic Wilson

Abstract
Computation on acyclic hypertrees or on related graphical structures has been discovered independently as an efficient method in several different fields, including probabilistic networks, statistics, relational databases, optimization and numerical mathematics. A generic algebraic structure, called a valuation algebra, has been identified as the underlying reason for the possibility of local computation. This structure unifies these local computation methods and makes them accessible for many further formalisms for representing information and knowledge; in particular, many kinds of logics, including non-classical ones, constraints, including soft constraints, different uncertainty calculi, including possibility theory and belief functions. This tutorial will provide a concise introduction into the underlying algebraic structure of valuation algebra, the corresponding architectures for local computation and gives some illustrative examples. This will allow the attendees to identify instances of problems for which local computation is a potential solution and give them the basic knowledge to select an appropriate architecture for the solution to their problems.

T02

Qualitative Spatial and Temporal Reasoning: from theory to applications.

CANCELLED

Abstract
Most of the knowledge used in real life is time and space dependent. The domain of qualitative spatial and temporal reasoning has evolved in AI as a response to the problem of reasoning about space and time when precise quantitative information is unavailable or impracticable. In the last few years, some general threads have emerged, which make an integrated presentation of the basic results feasible in a systematic way. Moreover, the range of applications is widening.
The tutorial introduces qualitative methods in spatial and temporal reasoning. It first presents some of the basic temporal formalisms for various models of time, including non-linear ones, and examines complexity issues in reasoning.  It then describes tractability results in temporal and spatial calculi such as Allen’s calculus and extensions of it, based on geometric and on syntactic methods. Similar questions are examined for the RCC calculi. Fuzzy extensions of the temporal and spatial calculi are discussed. The tutorial also describes some of the main applications of temporal and spatial reasoning (planning, diagnosis, natural language understanding, geographic information systems), as well as recent applications to the semantic web.

T03

Argumentation and Dialogue in AI.

Abstract
In many areas of human reasoning a formal demonstration of the truth of a proposition or the correctness of a plan is not possible. In such domains proof is not possible, and so instead rationality must be based on some more or less persuasive argument. Argumentation has long been of interest to informal logicians, but increasingly attempts are being made to make the notions involved sufficiently precise to be amenable to embodiment in computer systems. In particular the notion of dialogue can be modelled using argumentation, which is an important aspect of agent to agent communication. In this tutorial we will look at the characteristics of argument that distinguish it from proof; at ways in which a formal framework for reasoning about systems of argument can be provided; at representations of the structure of individual arguments; and at some example applications of these techniques.

T04

Specification and verification of epistemic  properties in multi-agent systems.

[Full-day]

CANCELLED

Abstract
This 8hr tutorial aims at introducing current state-of-the-art techniques and results in specification and verification of multiagent systems by means of epistemic logic. The intended target is the AI researcher interested in the relationship between AI and logic, and AI and verification of distributed systems. The tutorial is divided into two parts. In the first the emphasis is on the specification of AI systems by means of epistemic logic. In the second, the emphasis is on verification.

T05

Ontological Engineering.

Abstract
Ontologies provide a common vocabulary of an area and define, with different levels of formality, the meaning of the terms and the relationships between them. Ontological Engineering refers to the set of activities that concern the ontology development process,
the ontology life cycle, the methods and methodologies for building ontologies, and the tool suites and languages that support them. During the last decade, increasing attention has been focused on ontologies. Ontologies are now widely used in Knowledge Engineering, Artificial Intelligence and Computer Science; in applications related to areas such as Knowledge Management, Natural Language Processing, e-Commerce, Intelligent Information Integration, Bio-Informatics, Education; and in new emerging fields like the Semantic Web. This tutorial will present the theoretical foundations of Ontological Engineering, will describe the most outstanding ontologies that are currently available, and will cover the practical aspects of selecting and applying methodologies, languages, and tools for building ontologies. This tutorial will also aim at presenting commercial-oriented and research-oriented ontology-based applications. [more]

T06

Constraint Processing.

Abstract
This tutorial will provide participants with a firm understanding of the essential techniques for constraint processing. Presenters will summarize the huge amount of results published in the last years in a few lines of research, allowing attenders to get a clear and structured view of the field (see syllabus). The presentation will be focused on the algorithmic methods for constraint processing, without forgetting theoretical aspects and successful constraint applications. Algorithms will be described in terms of search, inference and hybrids. The tutorial will cover both types of constraint processing: hard constraints (classical constraint satisfaction) and soft constraints (constrained optimization). Modelling guidelines and environments for constraint processing will be discussed. To know more after the tutorial, a detailed list of references will be presented.

T07

Stochastic Local Search: Foundations and Applications.

CANCELLED

Abstract

T08

DataMining: From rule learning to relational subgroup discovery.

Abstract
Data Mining and Knowledge Discovery are concerned with the automated analysis of large-scale data in order to extract interesting patterns. This tutorial provides an overview of data mining and knowledge discovery from data stored in single tables (propositional learning) as well as multi-relational databases (relational learning), focusing on symbolic data mining techniques (decision trees and rules, association rules, inductive logic programming and relational data mining, etc.). Participants will obtain working knowledge of data mining, with an emphasis on commonly used techniques illustrated by examples. Evaluation and visualisation will also be discussed. This tutorial will be of interest to researchers, information technology professionals, information systems developers and managers, data analysts and institutional decision makers, and anyone else interested in applying intelligent data analysis methods to extract useful knowledge from data.

T09

Information Extraction: Theory and Practice.

CANCELLED

Abstract
The information age has made it easy to store large amounts of data. The proliferation of documents available on the Web, on corporate intranets, on news wires, and elsewhere is overwhelming. However, while the amount of data available to us is constantly increasing, our ability to absorb and process this information remains constant. Search engines only exacerbate the problem by making more and more documents available in a matter of a few key strokes. We will contrast the two major approaches to Information Extraction, the rule based approach, and the machine learning approach. In particular we will focus on comparing IE rule languages, and the various machine learning approaches. We will discuss the suitability of the two approaches to the tasks of Named Entity Recognition, Fact Extraction, and Relationship Extraction. We will provide an experimental evaluation of the two approaches and discuss way to build a third hybrid approach that can benefit from the strengths of the two approaches. We will demonstrate several real world applications of Information Extraction. We will present a general architecture of information extraction systems and will outline the algorithms and data structures behind the systems. The Tutorial will cover the state of the art in this rapidly growing area of research. Several real world applications of information extraction will be presented.

T10

Semantic Web : Vision, Tools and Applications.

Abstract
AI technology was never closer to changing the world as it is now that people like Tim Berners-Lee promote the “Semantic Web” as the next step in the WWW’s evolution. Machine-processable metadata and Web content as the basis for really “intelligent” personal information agents, for question answering and information processing on the Web instead of more or less precise document retrieval, making full use of Web infrastructures for distributed cross-organizational workflows and meaningful Enterprise Application Integration, these are the promises that people associate with the Semantic Web idea. However, though the term has been coined already a couple of years ago, the convincing commercial applications are still missing, the broad public adoption of the idea has not yet been taken place, and probably most of the involved technical problems have not even been articulated aloud (like integration of contradictory information, inferences with missing information, distributed inferences, processing of trust, knowledge representation for “normal end users”, or large-scale effects of non-monotonicity). Hence the first cautious warning cries for the next “AI winter” can already been heard.
In our tutorial we want to give a complete and thorough overview of today’s Semantic Web technology, its vision, technological basis, prospects, critical success factors and missing research, in order to give Semantic Web newcomers and interested AI researchers the chance to assess the potential of this new, challenging and thrilling application area. In order to do so, we will try to sketch the “whole picture”, i.e. describe the Semantic Web vision and visionary application scenarios, as well as already running example solutions and possibly interesting application scenarios, the complete technological basis, including ontology engineering, semantic annotation, and Semantic Web inference engines, and finally, missing links and required future research topics.

T11

Action Programming Languages and Cognitive Robotics.

Abstract
Based on 40 years of research in knowledge representation, action programming languages have recently been developed to endow software agents and autonomous robots with the ability to reason about their actions. These languages do by no means attempt to provide an alternative to sub-symbolic, probabilistic methods for robot control;  on the contrary, they are a complementary supplement: Thanks to a high level of abstraction, symbolic reasoning allows to solve complex tasks with huge state spaces by high-level programs which are easy to write, understand, and maintain and which allow one to implement flexible planning and acting strategies. Highly optimized implementations have recently been developed for various action programming languages. This tutorial will give an introduction to selected languages and systems. Participants will learn how to specify domains and how to write programs for endowing autonomous agents with problem solving abilities. The tutorial will provide an insight into the underlying mathematics and into the advantages and disadvantages of these languages in comparison. Several successful applications of these languages in combination with low-level control of autonomous robots will be discussed.