ESSLLI 2004 - Evening lectures

First Week: 9 - 13th of August, 2004
Date: Tuesday, August 10th
	Time: 21hs. Speaker: Vladimir Lifschitz Title: What Is Answer Set Programming? Abstract: Answer Set Programming (ASP) is a new form of declarative programming, oriented towards difficult combinatorial search problems. It has been applied, for instance, to plan generation and product configuration problems in Artificial Intelligence and to graph-theoretic problems arising in VLSI design and in historical linguistics. ASP helps organizers of computer science conferences assign submissions to referees in accordance with their preferences. Syntactically, ASP programs look like Prolog programs, but the computational mechanisms used in ASP are different: they are based on the ideas that have led to the development of fast satisfiability solvers for propositional logic. Material: Slides
Date: Thursday, August 12th
	Time: 20.30hs. Title: FoLLI General Meeting. Agenda: To be announced.
	Time: 21hs. Speaker: Karen Spärck Jones Title: Language and information processing: numbers that count Abstract: In early research on automated language and information processing (LIP), statistical approaches were advocated both for theoretical and for practical reasons. But though some innovative research was done, these approaches fell into disfavour except in the area of text retrieval and, later, speech recognition. In the last decade, however, statistical approaches have come into their own and are playing an important, even central, role in LIP. The lecture will review the development of these ideas and methods, consider their current contribution to LIP, and comment on the general lessons for LIP that have emerged from this research. Material: Slides References
Second Week: 16 - 20th of August, 2004
Date: Thursday, August 17th
	Time: 21hs. Speaker: Elisabeth André Title: Enhancing Embodied Conversational Agents with Social Intelligence Abstract: The last decade has seen a general trend in Human Computer Interaction to emulate certain aspects of human-human communication. Computers are ever less viewed as tools and ever more as partners or assistants to whom tasks may be delegated. In order to build dialogue systems that are able to communicate with the user in a more natural manner, the consideration of social aspects is inevitable. One aspect of social interaction is the use of politeness strategies as they are described in detail in Brown and Levinson's (1987) seminal work. Previous work has concentrated for the most part on the linguistic aspects of politeness strategies. But behaviours of politeness are inherently multimodal. For instance, friendly facial expressions may help to mitigate face threats evoked by criticism. In my talk, I will report on both verbal and nonverbal aspects of social communication and describe our approach to the development of computer-based embodied conversational agents that are able to vary their politeness behaviours based upon the emotional state of synthetic and human dialogue partners.
Date: Thursday, August 19th
	Time: 21hs. Speaker: Gérard Huet Title: Splitting the Entropy Gordian Knot Abstract: A natural language analyser must fight the entropy of the language combinatorics in order not to over-generate too much in its syntactic layers. A divide and conquer strategy is on the one hand to store a maximum of information in the lexicon, by lexicalizing the grammar, and by processing local feature constraints, and on the other hand to build a stochastic analyser, able to predict dislocations and to catch long distance dependencies. This general partition of information suggests two structures for treebank corpuses, one for lexicon acquisition, the other one for stochastic grammar training. It is argued that such general principles are useful for designing linguistic resources interoperable between computational platforms.

Further details will be announced here shortly.

2004-09-07