Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
CAREER: Scaling Up First-Order Logical Reasoning with Graphical Structure Eyal Amir, University of Illinois at Urbana-Champaign, eyal@cs.uiuc.edu
 

Summary: CAREER: Scaling Up First-Order Logical Reasoning with Graphical Structure
Eyal Amir, University of Illinois at Urbana-Champaign, eyal@cs.uiuc.edu
Project Summary
Criterion 1: What is the intellectual merit and quality of the proposed activity? The ability to represent and
reason about objects and relations between them is central to many approaches and applications in Artificial Intelli-
gence (AI), including common-sense query answering, natural-language processing, planning, and diagnosis problem
solving. In recent years the number of objects and relations that applications need to consider has increased dramati-
cally, and current real-world applications require reasoning mechanisms that can scale to thousands and more objects
and relations. Traditional approaches to logical reasoning that focus on propositional theories are impractical for such
real-world domain because propositional representations of the associated theories have explosive sizes, rendering
inference useless. On the other hand, current inference in First-Order Logic (FOL) is impractical here as well because
it focuses on mathematical theories that are much smaller and lack the structure of real-world common-sense domain
theories.
This proposal outlines a challenging career development plan that focuses on scaling up inference over FOL with
graph-based structures that are available in real-world domains. In this research the PI will focus first on inference in
FOL, and then apply the insights he gains from general FOL to inference with first-order structure on more specialized
scenarios (most times avoiding general FOL inference while still using the same structure). The PI is uniquely qualified
to do this because of (1) his work on speeding up FOL and propositional inference using partitions of predicate symbols
[AM00, AM05], (2) his work on AI-planning using partitions of actions [AE03], (3) his research on optimization
algorithms [Ami01, AKR03], and (4) his work on relational probabilistic methods in AI [dSBAR05].

  

Source: Amir, Eyal - Department of Computer Science, University of Illinois at Urbana-Champaign

 

Collections: Computer Technologies and Information Sciences