KLAP Seminar Home Page
Fall 2005
This is a weekly seminar which discusses issues related to
the areas of logic and constraint programming, knowledge representation,
and parallel processing.
It is organized by members of the
The Laboratory for Logic, DBs, and Advanced Programming.
KLAP stands for
Knowledge representation, Logic, and Advanced Programming.
The seminar is open to everyone.
Time: Monday, 11 am
Location: SH 118B
Upcoming Seminar
November 28
Hung Viet Le
spoke on
Answer Set Programming with Clause Learning
by Jeffrey Ward (The Ohio State University) and John S. Schlipf (University of Cincinnati)
Abstract
A conflict clause represents a backtracking solvers analysis of why a
conflict occurred. This analysis can be used to further prune the search
space and to direct the search heuristic. The use of such clauses has been
very important in improving the efficiency of satisfiability (SAT) solvers
over the past few years, especially on structured problems coming from
applications. We describe how we have adapted conflict clause techniques for
use in the answer set solver Smodels. We experimentally compare the
resulting program to the original Smodels program. We also compare to ASSAT
and Cmodels, which take a different approach to adding clauses to constrain
an answer set search.
Preliminary Schedule of talks
|
Date | Speaker | Title
|
|---|
| October 10
| Omar El Khatib
| Justification and Debugging of Answer Set Programs in ASP-Prolog
|
| October 17
| Chongbing Liu
| Nonmonotonic Inductive Logic Programming (NMILP)
|
| October 24
| Chongbing Liu
| Designing of a Nonmonotonic Inductive Logic Programming System
|
| October 31
| Tu Phan
| A new approach to conformant planning
|
| November 7
| Islam Elkabani
| Smodels^A - A System for Computing Answer Sets of Logic
Programs with Aggregates
|
| November 14
| Son Cao Tran
| What am I doing?
|
| November 21
| Thanksgiving holiday
|
| November 28
| Hung Viet Le
| Answer Set Programming with Clause Learning
|
Previous Seminars
November 28
Hung Viet Le
spoke on
Answer Set Programming with Clause Learning
by Jeffrey Ward (The Ohio State University) and John S. Schlipf (University of Cincinnati)
(
Paper in .pdf and presentation
slides)
Abstract
A conflict clause represents a backtracking solvers analysis of why a
conflict occurred. This analysis can be used to further prune the search
space and to direct the search heuristic. The use of such clauses has been
very important in improving the efficiency of satisfiability (SAT) solvers
over the past few years, especially on structured problems coming from
applications. We describe how we have adapted conflict clause techniques for
use in the answer set solver Smodels. We experimentally compare the
resulting program to the original Smodels program. We also compare to ASSAT
and Cmodels, which take a different approach to adding clauses to constrain
an answer set search.
November 14
Son Cao Tran
spoke on
his research
(Presentation
slides)
November 7
Islam Elkabani
spoke on
Smodels^A - A System for Computing Answer Sets of Logic
Programs with Aggregates
(Presentation
slides)
Abstract
We earlier presented a system called ASP-CLP for computing
answer sets of logic programs with aggregates. The implementation of ASP-CLP
relies on the use of an external constraint solver to deal with aggregate
literals and requires some modifications to the answer set solver. ASP-CLP
is based on a semantics that does not guarantee minimality of answer sets.
Furthermore, our experiments with ASP-CLP indicate that the cost of
communication between the constraint solver and the answer set solver proves
to be significant in large instances. In this work we explore an alternative
to ASP-CLP and develop a new system for computing answer sets of logic
programs with aggregates called Smodels^A. Smodels^A implements a different
translational semantics. This semantics does not impose any syntactic
restrictions on aggregates and at the same time does not produce non-minimal
answer sets. It naturally extends the traditional answer set semantics.
Moreover, the system can be easily implemented using off-the-shelf answer
set solvers.
October 31
Tu Phan
spoke on
A new approach to conformant planning
(Presentation
slides)
Abstract
Most of existing conformant planners search in belief state space.
However, one of the major disadvantages of this approach is that the
number of belief state is huge, double exponential in the number of
fluents. An alternative approach is to use approximation semantics.
Although significantly reducing the size of search space, approximation
semantics are criticized for its incompleteness.
In this talk, we will present a new approach, based on approximation
semantics. The approach is proven to be sound and complete with the
possible world semantics. Central in this approach is the concept of
"need-to-know" fluents. Intuitively, a need-to-know fluent is the one
whose truth values should be considered in the beginning in order to
guarantee that the approximation is complete. We also provide a polynomial
time algorithm to determine such fluents.
October 24
Chongbing Liu
spoke on
Designing of a Nonmonotonic Inductive Logic Programming System
(Presentation
slides)
Abstract
In this talk the problem of designing an inductive logic
learning system under nonmonotonic logic framework, is discussed.
First, we summarize the basic algorithms proposed by Chiaka Sakama
for learning from a single positive example, from a single negative
example, and from a set of examples. we also illustrate their
properties and restrictions. Second, we show the pros and cons of
several possible sequential designs. Issues include the learning
modes (i.e., incremental or batch learning) and search strategies
(i.e.,top-down or bottom-up search). Lastly, we look into the possible
ways of paralleling the learning processing under the nonmonotonic
logic framework.
October 17
Chongbing Liu
spoke on
Nonmonotonic Inductive Logic Programming (NMILP)
(Presentation
slides)
Abstract
Most of the existing Inductive Logic Programming (ILP) systems are based
on classical Horn logic. However, this language is not sufficiently
expressive for representing and reasoning with incomplete human
knowledge.
In this talk, I will present people's work on inductive learning under
nonmonotonic logic framework, where both the background knowledge and the
hypotheses may contain negations. First, I will show two approaches,
Closed World Specialization and Complete Inverse Entailment which
represent effort for extending traditional ILP towards nonmonotonic ILP.
These approaches are based on SLDNF-resolution. Next, I will introduce
approaches based on stable model semantics, mainly including Chiaki
Sakama's algorithms to learn a rule from a single positive example, a
single negative example, and to learn rules from a set of examples.
October 10
Omar El Khatib
spoke on
Justification and Debugging of Answer Set Programs
in ASP-Prolog.
(Presentation
slides)
Abstract
The paper extends the concept of justification to the context of
Answer Set Programming ---a recent paradigm that builds on
the foundations of logic programming, answer set semantics, and
non-monotonic reasoning. A justification describes the
support for the truth value of each atom in an answer
set of a logic program, and it can be employed as
a tool for reasoning and debugging of answer set programs.
The paper describes the implementation of the notion of justification
in the ASP-Prolog system along with some examples of its applications.
KLAP seminar in the Fall 2003.
This page is maintained by
Inna Pivkina
