Chairs

Vice-Chairs

Members

Paul Cohen
University of Arizona, USA

What shall we do next? The challenges of AI midway through its first century

Invited speakers:
 

Hendrik Blockeel
K.U.Leuven , Belgium, and Leiden University, The Netherlands

Exposing the causal structure of processes by learning CP-logic programs

This talk starts with a brief introduction to probabilistic logic learning, after which we will focus on a relatively new formalism known as CP-logic. CP-logic stands for "causal probabilistic logic". It is a knowledge representation formalism that allows us to write down rules that indicate that a certain combination of conditions may cause certain effects with a particular probability (e.g., tossing a coin may cause a result of heads or tails, each with 50% probability). Besides the fact that this formalism is interesting for knowledge representation in itself, it also offers interesting opportunities from the machine learning point of view.  Indeed, given the semantics of these CP-logic programs, learning them from data amounts to extracting probabilistic causal influences from data.  We will discuss recent research on learning CP-logic programs, including: algorithms for learning them; how they relate to graphical models; and applications of learning CP-logic programs.

Building Structured Web Community Portals via Extraction, Integration, and Mass Collaboration

Abstract: The World-Wide Web hosts numerous communities, each focusing on a particular topic. As such communities proliferate, so do efforts to build community portals. Most current portals are organized according to topic taxonomies. Recently, however, there has been a growing effort to build structured data portals (e.g., IMDB, Citeseer) that present a unified view of entities and relationships in the community. Such portals can prove extremely valuable in a wide range of domains. But how can we build them efficiently?

In this talk, I will present a new research vision that addresses this question. The goal is to develop a system that a small team (or ideally just one person) can quickly deploy to build an initial (but already useful) structured portal, then leverage the entire community in a mass collaboration fashion to improve and expand this portal. As such, the research agenda requires combining and extending research in information extraction, information integration, and Web 2.0 technologies, among others. This agenda is actively being pursued in the Cimple project, a joint effort between the University of Wisconsin and Yahoo Research. In the talk I will describe recent progress in Cimple, portal prototypes, lessons learned, and future directions. I will focus in particular on how Cimple raises interesting and novel challenges for both AI and database research. More information about Cimple can be found at www.cs.wisc.edu/~anhai/projects/cimple.

Large Scale Corpus Analysis and Recent Applications

Abstract: Recent progress of corpus and machine learning-based natural language processing methodologies have made it possible to handle large scale corpus with a quite high accuracy.  The speaker is now involved in a project for constructing a large scale contemporary Japanese balanced corpus, aiming at constructing automatic annotation tools on various levels of natural language analyses.  I will first introduce our activities on corpus based natural language analyzers for word dependency parsing and anaphora resolution and annotated corpus management environment.  Then, I will explain recent natural language applications such as sentiment/opinion mining and knowledge extraction from a large scale text data like Weblogs.

Biography: Yuji Matsumoto received Master and PhD degrees in Information Science from Kyoto University.  He is currently a professor at the Graduate School of Information Sicence in Nara Institute of Science and Technology.  He and his group have developed lots of machine learning-based natural language processing tools such as ChaSen and CaboCha as well as Japanese and Chinese dictionaries, all of which are publicly available.  His research interests cover corpus-based natural language processing, knowledge acquisition, lexical semantics, text mining and integration of empirical and theoretical natural language techniques.

 
 Asian Office of Aerospace Research and Development AFOSR/AOARD