------------------------------------------------------------------------
Speaker:        Dr. Jerome LANG
                IRIT (CNRS et Universite Paul Sabatier)
                Toulouse, France
                lang@irit.fr

Topic:          "From Belief Update To Reasoning About Action"

Date:           Monday, 3 December 2001

Time:           4:00pm - 5:00pm

Venue:          Lecture Theater F
                (Leung Yat Sing Lecture Theater, near lift nos. 25/26)
                HKUST

ABSTRACT:

Belief update is a form of belief change that aims at representing an
explicit evolution of the world. We first give some background on belief
update and its complexity. Then we investigate the issue of drawing action
representation formalisms from belief update operators.  We start by
giving a new family of dependence-based update operators that are
well-suited to the representation of simple actions and that are not more
complex than pure propositional logic. Then we show that belief update
(whatever the operator chosen) fails to represent in a satisfactory way
conditional actions, nondeterministic actions, and concurrent actions.
This leads to define extended belief update operators where these three
issues are dealt with. We show that this gain in expressivity does not
induce a complexity gap. Then we turn to the issue of using these extended
update operators in a planning process. In this framework, we define
conditional plans and we express plan verification and plan existence in a
simple way; lastly we investigate the complexity of the latter two
problems under several observability assumptions.



**********************
Biography:

Jerome Lang is a researcher at the French National Center for Scientific
Research (CNRS) since 1991. He has obtained a PhD from Universite Paul
Sabatier in 1991. His past and/or current domains of research are:
computational complexity, nonmonotonic reasoning, reasoning under
uncertainty,  belief revision, reasoning under inconsistency, reasoning
about action, preference representation, planning under incomplete
knowledge, decision-theoretic models for AI.


                For enquiries, please call 2358 7008
                  **** ALL are Welcome ****
--------------------------------------------------------------------------