Workshop on Practical Commonsense

.

Topics of Interest

This workshop will bring together semantic web, representation, knowledge acquisition and reasoning approaches that are aimed at broad coverage, vastly incomplete reasoning, with background knowledge, over enormous, globally inconsistent and partially incorrect or inexact knowledge bases. Areas of interest include:

- Large scale probabilistic and contextual reasoning,
- Reasoning with various knowledge representations: logic, logical form, association patterns, natural language,
- Automated knowledge acquisition, particularly for broad domain background knowledge,
- Partially correct inference mechanisms,
- Fast, scalable but incomplete inference mechanisms,
- Analysis of applications of commonsense (broad coverage, incomplete, rapid inference with background knowledge),
- and related topics. 

Scope

Reasoning and knowledge representation (KR), along with learning, perception and expression, have lain at the core of Artificial Intelligence research since the creation of the field. The earliest representations and reasoning systems were effective in the particular domains to which they applied, but only in those domains. They were quickly supplanted following the development of effective algorithms for first order inference, and much of the subsequent work on KR and automated reasoning has focused on theoretical issues in producing predicate calculus-based knowledge representations and in inference over these representations. Although not universally the case, much of this work has concentrated on representations and methods for exact, complete inference over relatively small, globally consistent knowledge bases. Recent technological and social trends are pushing the world in quite another direction. Huge amounts of material (e.g. Wikipedia, Wikihow, online media) produced by large numbers of relevant people (e.g. LinkedIn, the worldwide scientific literature, mySpace) are now both relevant and available. This new availability confronts us with material overwhelming our natural information processing ability. And the material is heterogeneous, imprecise, and enormous. Complete inference over such material is both hopeless and pointless; what is needed is a comprehensive ability to do something useful with the material, that wouldn't otherwise have been done. This is commonsense reasoning. 

Organising Committee

Michael Witbrock (Cycorp) Este endereço de email está protegido contra spam bots, pelo que o Javascript terá de estar activado para poder visualizar o endereço de email

Hugo Liu (MIT Media Lab) Este endereço de email está protegido contra spam bots, pelo que o Javascript terá de estar activado para poder visualizar o endereço de email

Dunja Mladenic (Institut Josef Stefan) Este endereço de email está protegido contra spam bots, pelo que o Javascript terá de estar activado para poder visualizar o endereço de email

Kenneth Forbus (Northwestern University) Este endereço de email está protegido contra spam bots, pelo que o Javascript terá de estar activado para poder visualizar o endereço de email

Paulo Urbano (Universidade de Lisboa) Este endereço de email está protegido contra spam bots, pelo que o Javascript terá de estar activado para poder visualizar o endereço de email (local arrangements)

Program Committee

Chris Welty, IBM research, US
John Davies, British Telecom, UK
Clive Best, EU JRC Ispra, Italy
Henry Lieberman, MIT, USA
Catherine Havasi, MIT, USA
Larry Birnbaum, Northwestern U, USA
Marko Grobelnik, IJS, Slovenia
Peter Haase, Uni Karlsruhe, Germany
Francesca Lisa, Univ di Bari, Italy
Miguel-Angel Sicilia, Univ de Alcala, Spain