Speaker: Prof. Jia-Huai You (University of Alberta, Canada)
Abductive reasoning subscribes to reasoning processes
where explanatory hypotheses are formed and evaluated
with respect to a knowledge base and an observation.
Many intelligent tasks, including diagnosis, scientific
discovery, legal reasoning, and natural language
understanding, have been characterized as abduction.
Logic programming offers an attractive approach to abductive
reasoning, partially due to its efficient top-down reasoning
mode. Logic programs are extended to allow negation in rule
bodies. While this addition has substantially increased the
expressive power of logic programs, it also caused a problem
for proof procedures; namely the general difficulty in
processing negative queries with variables. This is the well-known
problem of "problematic quantification patterns."
In this talk, we present an abductive proof framework, where all
programs and queries are processed under a uniform representation.
We show that the quantification problem is essentially a problem
of solving quantified equations and disequations. Provided that
such a reasoner is available, the proposed framework is sound and
complete with respect to the predicate completion semantics.
About the Speaker:
Jia-Huai You obtained his PhD from the University of Utah in 1985, and
first served on the faculty of Rice University, and then joined the Department
of Computing Science at the University of Alberta, where he is now a Professor.
His research interests are in logic-based Knowledge Representation and Reasoning,
Constraint Satisfaction, Nonmonotonic Logics, and various programming languages for
declarative problem solving. Dr. You has published substantially in first-rate
international journals and conferences, such as Artificial Intelligence, JCSS,
ACM TOCL, JAR, JLP, TCS, IJCAI, PODS, POPL, ICLP, etc. He has served on the PC
of a number of conferences, including IJCAI, AAAI, ECAI, LPNMR, etc. His most recent
research interest lies on issues related to integration of SAT, CSP and Answer Set