Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 093
AIList Digest Thursday, 2 Apr 1987 Volume 5 : Issue 93
Today's Topics:
Seminars - AI, Mathematical Programming, and VLSI Design (Rutgers) &
Automating Theory Formation (Rutgers) &
Concept Learning (Ames) &
Argo: Analogical Reasoning for Design Problems (Rutgers) &
Decomposition for Hierarchical Problem Solving (Rutgers),
Conferences - Simulation & Protocol Specification
----------------------------------------------------------------------
Date: Wed, 25 Mar 87 18:17:55 EST
From: liew@aramis.rutgers.edu (Liew)
Subject: Seminar - AI, Mathematical Programming, and VLSI Design (Rutgers)
The next design colloquim will be held on Thursday (march 26th) at
1:30pm in TCB 103. Most of you are unfamiliar with the location of
TCB 103 so we will meet at Hill 423 at 1:15 and proceed from there.
The speaker will be Wayne Wolf of ATT Bell Laboratories and the title
of his talk is "Artificial Intelligence, Mathematical Programming and
VLSI Design". The suggested readings are:
Wolf, Kowalski, McFarland "Knowledge Engineering Issues in VLSI
Synthesis", AAAI-86.
Brayton, et al., "Multiple-Level Logic Optimization System",
ICCAD-86, pp. 356-360.
Gregory, et al., "SOCRATES: A System for Automatically Synthesizing
and Optimizing Combinational Logic", DAC-86, pp. 79-85.
Shin and Sangiovanni-Vincentelli, "MIGHTY: A Rip-Up and Reroute
Detailed Router", ICCAD-86, pp. 2-5.
Joobani, "WEAVER: A Knowledge-Based Routing Expert", PhD
dissertation, CMU, 1985.
----------------------------------------------------------------------
Abstract:
Title: Artificial Intelligence, Mathematical Programming, and VLSI Design
Speaker: Wayne Wolf, AT&T Bell Laboratories, Murray Hill
Artifical intelligence techniques have found their
greatest success in diagnosis and classification problems. The
application of AI to design problems is relatively new. In
this talk I want to consider how the intellectual tools that
AI brings to the design problem can best be used by contrasting
two paradigms: artificial intelligence and mathematical programming.
I will argue that mathematical programming is a more powerful
paradigm than AI for a lot of synthesis problems because
mathematical programming a) allows better application of
brute force; b) encourages us to formulate solvable problems.
I will argue that AI is a more powerful paradigm for
knowledge representation because it provides a lot of tools
for separating particular pieces of knowledge from the
engines used to maintain them.
The talk will be in three parts:
1) The VLSI design problem: what is hard about VLSI
design; what tools people need to make bigger, better designs;
what people would do with VLSI synthesis if they had it.
2) Synthesis and search: search in AI and mathematical
programming; problem formulation and search; results in
application of AI and mathematical programming techniques
to some design problems.
3) Synthesis and knowledge representation: why
knowledge representation is important; examples of KR
problems and solutions from Fred, the database; how
AI knowledge representation and mathematical programming
complement each other in Lucy, the controller designer.
------------------------------
Date: 26 Mar 87 17:01:17 EST
From: SOO@RED.RUTGERS.EDU
Subject: Seminar - Automating Theory Formation (Rutgers)
THE III, AN INFORMAL SEMINAR FOR AND BY STUDENTS,
--- INITIATES ITS SPRING SEASON ---
Title: Automating Theory Formation --
Postulation of Enzyme Kinetic Models
and Experimental Design
Date: April 7th, Tuesday
Time: 11:00 AM
Place: Hill 423
Speaker: Von-Wun Soo
This is a practice talk for my Ph. D. thesis defense. I would like to
present the work that I have been involved for the past five years.
I cordially invite you to come, support, and make comments before
my final defense.
Abstract:
In this talk I discuss how expert reasoning in scientific research such
as designing biochemical experiments or postulating kinetic mechanisms can
be modeled. Broadly speaking, designing an experiment, an important compoent
of scientific theory formation, can be viewed as a process of searching
and testing plausible decompositions of a hypothesis space.
In my thesis, I show how the results of qualitative reasoning and a
set partition method can be used to select experimental setups that
discriminate a set of plausible models. The interpretation of
experimental results, the critiques of previous experiments,
and comparisons of similarities and discrepancies among experiments
are all related issues that lead us to the automation of
scientific discovery.
------------------------------
Date: Thu, 26 Mar 87 21:31:07 PST
From: SIMS%PLU@ames-io.ARPA
Subject: Seminar - Concept Learning (Ames)
Title: LEARNING CONCEPTS TO IMPROVE PERFORMANCE:
The Role of Context
By: Dr. Richard Keller
(KELLER@RED.RUTGERS.EDU)
Computer Science Department
Rutgers University
Where: NASA AMES
When: Monday, April 6
Concept learning, like most intelligent behavior, should be
influenced by the context in which the behavior takes place. If
concept learning occurs in the context of improving the performance
of a problem solving system, then the type of concept learned and
the form of its description should depend on the goals and the
capabilities of the problem solver. Unfortunately, most current
inductive learning systems incorporate a set of fixed, implicit
assumptions about the problem solver being improved by learning.
This causes problems when the original problem solver changes over
time, and also makes it difficult to reuse the same inductive system
to improve a different problem solver.
As an alternative to the inductive framework, I describe a new
concept learning framework -- the concept operationalization
framework -- which makes contextual assumptions more explicit and
easier to change. To illustrate the new framework, I discuss how an
existing inductive system (the LEX system [Mitchell et al. 1981])
was converted to a concept operationalization system (the MetaLEX
system). In contrast with LEX, MetaLEX adapts more successfully to
certain changes in its learning context, learns contextually
suitable approximations of its target concept as necessary or
expedient, and has the potential to automatically generate its own
concept learning tasks to improve its problem solver.
------------------------------
Date: Mon, 30 Mar 87 15:22:26 EST
From: liew@aramis.rutgers.edu (Liew)
Subject: Seminar - Argo: Analogical Reasoning for Design Problems
(Rutgers)
There will be a design colloquim on Tuesday March 31st at 10:30 am in
Hill 423. The speaker will be Ramon Acosta of MCC and an abstract of
his talk is given below. A copy of his paper is in JoAnn Gabinelli's
office (Hill 408).
Argo: An Analogical Reasoning System for Solving Design Problems
Michael N. Huhns and Ramon D. Acosta
Microelectronics and Computer Technology Corporation
AI/KBS and VLSI CAD Programs
3500 West Balcones Center Drive
Austin, TX 78759
The static and predetermined capabilities of many knowledge-based design
systems prevent them from acquiring design experience for future use.
To overcome this limitation, techniques for reasoning and learning by
analogy that can aid the design process have been developed. These
techniques, along with a nonmonotonic reasoning capability, have been
incorporated into Argo, a tool for building knowledge-based systems.
Closely integrated into Argo's analogical reasoning facilities are
modules for the acquisition, storage, retrieval, evaluation, and
application of previous experience. Problem-solving experience is
acquired in the form of a problem-solving plan represented as a
rule-dependency graph. From this graph, Argo calculates a set of
macrorules, each based on an increasingly abstract version of the plan.
These macrorules are partially ordered according to an abstraction
relation for plans, from which the system can efficiently retrieve the
most specific plan applicable for solving a new problem. The use of
abstraction in a knowledge-based application of Argo allows the system
to solve problems that are not necessarily identical, but just analogous
to those it has solved previously. Experiments with an application for
designing VLSI digital circuits are yielding insights into how design
tools can improve their capabilities and extend their domains of
applicability as they are used.
------------------------------
Date: 31 Mar 87 10:36:22 EST
From: KAPLAN@RED.RUTGERS.EDU
Subject: Seminar - Decomposition for Hierarchical Problem Solving
(Rutgers)
PhD Oral Qualifying Examination for Mr. S. Mahadevan
Mr. Mahadevan's examination is scheduled for Wednesday, April 1 at 10:30 AM
in Hill 423. The examination committee is chaired by T. Mitchell, and includes
T. McCarty, J. Mostow, and L. Steinberg. DCS faculty are welcome to attend;
graduate students are invited to the public portion of the examination. Mr.
Mahadevan's dissertation proposal is abstracted below:
LEARNING DECOMPOSITION METHODS TO IMPROVE HIERARCHICAL
PROBLEM-SOLVING PERFORMANCE
Previous work in machine learning on improving problem-solving
performance has usually assumed a state-space or "flat"
problem-solving model. However, problem-solvers in complex domains,
such as design, usually employ a hierarchical or problem-reduction
strategy to avoid the combinatorial explosion of possible operator
sequences. Consequently, in order to apply machine learning to
complex domains, hierarchical problem-solvers that automatically
improve their performance need to designed. One general approach is
to design an interactive problem-solver -- a learning
apprentice -- that learns from the problem-solving activity of expert
users. In this talk we propose a technique, VBL, by which such a
system can learn new problem-reduction operators, or decomposition
methods, based on a verification of the correctness of example
decompositions. We also discuss two important limitations of the VBL
technique -- intractability of verification and specificity of
generalization -- and propose solutions to them. Finally, we present
a formalization of the problem of learning decomposition methods based
on viewing actions and problems as binary relations on states.
------------------------------
Date: Thu, 26 Mar 1987 19:05 CST
From: Leff (Southern Methodist University)
<E1AR0002%SMUVM1.BITNET@wiscvm.wisc.edu>
Subject: Conferences - Simulation & Protocol Specification
The Society for Computer Simulation Eastern Simulation Conferences
April 6-9, 1987, Orlando, Florida
AI and Simulation at Johnson Space Center, Robert Salvely (verbal presentation)
Flight Simulator Evaluation of Aircraft Systems Using AI Technology (verbal ..
Edward M. Huff, NASA Ames Research Center
An Expert System fo rManaging Multiple Cooperating Expert Systems (verbal ...
A. Gerstenfeld, Geoffrey Gosling, David S. Touretzky Worcester Polytechnic
The Simulation of Simple Analog and Discrete Circuits from a Knowledge Base
Representaiton of Structure and Function
NASA, Kennedy Space Center
Acknowledge2: A Knowledge Acquisition System
Pradip Dey, Kevin D. REilly, J. Todd Brown, University of Alabama at Birmingham
Knowledge Corpora Connectivities - Toward the Construction of a Thought
Simulator Testbed
Alhad M. Chande, Marti.n Marietta Baltimore Aerospace, Joe Clema, IIT
Research Institute
Qualitative Expert Systems: A Demographic Simulator with Heuristic Reasoning
Walt Conley, W. Lawrence, U. Sengupta, R. Hartley, M. Coombs, New Mexico
State University
Model Management in Knowledge Based Simulation
Hawa Singh, Alan Butcher, R. Reddy, West Virginia University
Constraint Directed Reasoning for Simulation Problem Formulation
Neena Sathi, Gary STrohm, Thomas Morton, Sean Winters, Carnegie Group Inc.
Knowledge-Based Resource Behavior
Allen Matsumoto, V. Baskaran, Beth Marvel, Carnegie Group
Sonar Plexus - Enhancing a Command and Control Simulation with Reasoning
Marc R. Halley, Thomas MIller, Craig Hougum, William Mosenthal
Analytic Sciences Corporation
Computer System Simulation in Scheme
Daniel B. Pliske The Analytical Sciences Corporation
Real Time Intelligent System Analysis by Dsicrete Event Simulation
J. M. Poole, T. M. McDermott, D. P. Glasson,
The Analytica Sciences Corporation
The Mobile Intercontinental Ballistic MIssible Simulation
Douglas Roberts, J. Darrell Morgeson, Jared S. Dreicer, Howard W. Egdorf
Los Alamos National Laboratory
SIMSMART: Dynamic Simulation fo rAutomated Control of Complex Industrial
Processes
Don Waye Applied High Technology Limited
Applicability of AI Techniques to Simulation Models
Norman R. Nielsen, SRI International, Victoria P. Gilbert, Intellicorp
Improving Effectivenes of Computer Simulation SModeling with Knowledge-
Based Problem-Solving Capability
Ronak Shodhan, J. J. Talavage, Purdue University
Expert Systems within Simulations
JohnPaul SanGiovanni, Jockey Holley Technologies
A Communication Network Model of the Brain
Ray Moses, Boeing Aerospace Company
An ARtificial Intelligence (AI) Simulation Based Approach for Aircraft
Maintenance Training
Lee Keskey, Dave Sykes, Honey Well Inc.
Knowledge Representation in Ada
Sumitra M. REddy, Francis L. VAn Scoy, West Virginia University
A Simulator of an Automatic Text Reading System
Nikolaos G. Bourbakis, George Mason University, Scott Schneider, IDA
Two-dimensional Image Scanning for Hierarchical Data Structures and Its
Simulation
Nikolaos G. Bourbakis, George Mason University
Cognitive Learning Theory: A Tool for Modelling and Simulation
Donald A. MacCuish, ICSD Corporation
A Computer Simulation Program of Animal Maze Learning
Roger Ingliss, Warren Marchioni, Montclair High School
+++++++++++++++++++++++++++++++++++++++++++++++++++++++
Protocol Specification, Testing and Verification: VII
May 5- 8, 1987, IFIP Protocol Symposium Interconventional Ltd.
c/o SWISSAIR CH-8058 Zurich-Airport, Switzerland
Communicati.ng Rule Systems
L. F. Mackert & I. Neumeier-Mackert
IBM European Network Center, Heidelberg
An Atomic Calculus of Communicati.ng Systems
L. Logrippo and A. Obaid
University of Ottawa
Fundamental Results for the Verification of Observational Equivalence:
A Survey
T. bolognesi, CNUCE, Pisa, S. A. Smolka, SUNY, STony Brook
Proof of Specification Properties by Using Finite State Machines and
Temporal Logic
A. R. Cavalli, F. Horn, CNET, Issy, Les Moulineaux
Translation of Formal Protocol Specifications to VLSI Designs
A. S. Krishankumar, B. Krishnamurthy, K. Sabnani
AT&T Bell Labs, Murray Hill
------------------------------
End of AIList Digest
********************
