Dresden University of Technology, Germany
GENERAL GAME PLAYING
A general game player is a program that accepts formal descriptions of arbitrary games and plays these games without human intervention. One of the grand AI challenges, General Game Playing requires to combine techniques from a wide range of areas including knowledge representation, automated reasoning, heuristic search, planning, and learning. This talk will provide insights into our system FLUXPLAYER, which has recently been crowned the world champion at the Second General Game Playing Competition at AAAI-06 in Boston.
University of Washington, U.S.A.
UNIFYING LOGICAL AND STATISTICAL AI
Intelligent agents must be able to handle the complexity and uncertainty of the real world. Logical AI has focused mainly on the former, and statistical AI on the latter. In recent years, the field of statistical relational learning has set about combining the two. In this talk, I will survey recent developments in this area, with a particular focus on my group's work on Markov logic. Markov logic combines logic and probability by attaching weights to first-order formulas and viewing them as templates for features of Markov networks. Inference algorithms for Markov logic draw on ideas from satisfiability, Markov chain Monte Carlo and knowledge-based model construction. Learning algorithms are based on the voted perceptron, pseudo-likelihood and inductive logic programming. Markov logic has been successfully applied to problems in entity resolution, link prediction, information extraction and others, and is the basis of the open-source Alchemy system. I will conclude by discussing open problems and research directions in statistical relational AI.
(Joint work with Stanley Kok, Daniel Lowd, Hoifung Poon, Matt Richardson, Parag Singla, and Marc Sumner.)
Jozef Stefan Institute, Ljubljana, Slovenia
CONTENTS AND COAUTHORSHIP ANALYSIS OF ILP PUBLICATIONS DATA
The ILPnet2 database is publicly available on the Web and contains information about publications in the area of Relational Data Mining and Inductive Logic Programming, published in the period 1971-2003. This talk presents the results of co-authorship analysis obtained using the Pajek network analysis program, and the analysis of the contents of ILPnet2 publications with OntoGen, a system for data-driven semi-automatied topic ontology construction. In addition to the results of exploratory data analysis with Pajek and OntoGen, the talk will present a recent advancement of the ontology construction method based on automated term extraction, and the visualizations of the ILPNet2 topics and the co-authorship network.
(Joint work with Blaz Fortuna and Miha Grcar)