Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 154
AIList Digest Monday, 23 Jun 1986 Volume 4 : Issue 154
Today's Topics:
Seminars - Motion Planning (GMR) &
Unifying Principles of Machine Learning (UPenn) &
Parallel Execution of Logic Programs (UTexas) &
Why Planning Isn't Rational (SRI) &
Symbolic Representation of Waveforms (CMU)
----------------------------------------------------------------------
Date: Mon, 16 Jun 86 15:26 EST
From: "Steven W. Holland" <HOLLAND%RCSMPA%gmr.com@CSNET-RELAY.ARPA>
Subject: Seminar - Motion Planning (GMR)
Seminar at General Motors Research Laboratories, Warren, Michigan (GMR):
Motion Planning: a Survey of the State of the Art
Joseph O'Rourke
Department of Computer Science
Johns Hopkins University
Monday, June 23, 1986
Abstract
Motion planning from the viewpoint of computational geometry is the problem
of moving an object (a robot hand, for example) from an initial to a final
position in the presence of fixed obstacles. A large number of algorithms
have been developed for various cases of this problem recently. I will
describe the two main paradigms for solving this problem, growing
algorithms and Voronoi diagram algorithms, and survey the known results.
Several special cases will be discussed, including moving a disk, moving a
ladder, moving through a door, and moving around a corner. I will also
touch on the more complex problem of moving through an environment which is
not itself fixed, for example, one that contains several independently
moving robots.
Joseph O'Rourke has been Assistant Professor at Johns Hopkins University
since receiving the Ph.D. degree in Computer Science at the University of
Pennsylvania in 1980. His dissertation research was in computer vision,
and he has published in pattern recognition, but now his research is
focused on computational geometry. O'Rourke is a NSF Presidential Young
Investigator.
-Steve Holland, Computer Science Department
------------------------------
Date: Wed, 18 Jun 86 11:08 EDT
From: Tim Finin <Tim%upenn.csnet@CSNET-RELAY.ARPA>
Subject: Seminar - Unifying Principles of Machine Learning (UPenn)
CIS Colloquium
3 p.m. Thursday, June 19, 1986
216 Moore School, University of Pennsylvania
MACHINE LEARNING: UNIFYING PRINCIPLES AND RECENT PROGRESS
Ryszard S. Michalski
University of Illinois
Machine learning, a field concerned with developing computational theories of
learning and constructing learning machines, is now one of the most active
research areas in artificial intelligence. An inference-based theory of
learning will be presented that unifies the basic learning strategies. Special
attention will be given to inductive learning strategies, which include
learning from examples and learning from observations and discovery.
We will show that inductive learning can be reviewed as a goal-oriented and
resource-constrained inference process. This process draws upon the learner's
background knowledge, and involves a novel type of inference rules, called
inductive inference rules. In contrast with truth-preserving deductive rules,
inductive rules are falsity-preserving.
Several projects conducted at our AI Laboratory at Illinois will be briefly
reviewed, and illustrated by examples from implemented programs.
------------------------------
Date: 19 Jun 86 17:14:14 GMT
From: ucbcad!nike!lll-crg!seismo!ut-sally!leung@ucbvax.berkeley.edu
(Clement Leung)
Subject: Seminar - Parallel Execution of Logic Programs (UTexas)
AN ABSTRACT MACHINE BASED EXECUTION MODEL FOR COMPUTER ARCHITECTURE DESIGN
AND EFFICIENT IMPLEMENTATION OF LOGIC PROGRAMS IN PARALLEL
Manuel V. Hermenegildo
Dissertation Defense
The University of Texas at Austin
Department of Electrical and Computer Engineering
June 20, 1986 - 11:00 am - ENS431
Parallel execution represents one way in which the execution speed of logic
programs can be increased beyond the limits of conventional systems.
However, most proposed parallel logic programming systems lack the
optimizations and storage efficiency of high-performance sequential
implementations.
A parallel execution model for logic programs will be presented which is
based on extending to a parallel environment the techniques introduced by
the "Warren Abstract Machine", which have already made very fast and space
efficient sequential systems a reality. Therefore, the model is capable of
retaining sequential execution speed similar to that of current high
performance systems, while extracting additional gains by efficiently
supporting parallel execution. The model is described down to the Abstract
Machine level, specifying data areas, operation, and a suitable instruction
set. Several techniques are introduced which offer efficient solutions to
areas of parallel Logic Programming implementation previously considered
problematic or a source of considerable overhead, such as the specification
of control and management of the execution tree, the detection and handling
of variable binding conflicts in AND-Parallelism, support for "don't know"
non-determinism and treatment of distributed backtracking, goal scheduling
and memory management issues etc. These claims are supported by simulations.
------------------------------
Date: Thu 19 Jun 86 17:22:52-PDT
From: Amy Lansky <LANSKY@SRI-AI.ARPA>
Subject: Seminar - Why Planning Isn't Rational (SRI)
WHY PLANNING ISN'T RATIONAL
Terry Winograd (TW@SAIL)
Stanford University
(Computer Science, Linguistics, and CSLI)
11:00 AM, MONDAY, June 23
SRI International, Building E, Room EK242 (note room change)
Orthodox AI approaches to describing and achieving intelligent action
are based on a "rationalistic" tradition in which the focus is on a
process of deducing (using a representation of some kind) the
consequences of specific acts (operations) and searching for a sequence
of acts that will lead to a desired result (goal). This works
reasonably well for some limited domains, but falls far short of being a
general theory of intelligent action. It does not work well in the
small (how I operate my finger muscles, or where an amoeba slithers), or
in the large (how I conduct my life or where my research is headed).
Even in the cases of clearly explicit rational planning (e.g. planning a
bank robbery), the relation between plan and execution is not easy to
capture (what happens when the teller sneezes?).
In a recent book written jointly with Fernando Flores, I have proposed a
different basis for looking at action and cognition, focussing on the
"thrownness" of action without reflection, and on the open-endedness of
interpretation. Any alternative such as ours must address several
obvious questions:
Why is the naive view of rational decision-making and action so
intuitively plausible if it isn't right?
How can we account for the evolution of complex behavior which is
effective in an environment?
What implications does it have for AI and the design of computer
systems in general?
I will address these questions and related others, focussing on some
different issues from those raised in my talk to CSLI a couple of weeks
ago on "Why language isn't information".
VISITORS: Please arrive 5 minutes early so that you can be escorted up
from the E-building receptionist's desk. Thanks!
------------------------------
Date: 18 Jun 86 18:59:25 EDT
From: Dave.McKeown@maps.cs.cmu.edu
Subject: Seminar - Symbolic Representation of Waveforms (CMU)
Monday June 23 1:30pm 5409 Wean Hall
A Symbolic Representation of Waveforms Using Multi-Resolution Analysis
Dr. Aviad Zlotnick
Department of Mathematics and Computer Science
Hebrew University, Israel
A multi-resolution technique for ``qualitative'' analysis of waveforms was
suggested by Witkin in 1983, and has since been studied extensively, both
in theory and in practice. In the first part of the talk we reconsider
Witkin's definition of qualitativity and outline a few weaknesses of his
method. In the second part we describe a representation based on an
alternative definition of qualitativity. We show that our method results
in waveform descriptions which are nearer to human intuition, are easier
to compute and can incorporate more domain knowledge. Furthermore, a
symbolic (verbal) description of waveforms derived from this representation
is shown to capture the waveforms' essential visual properties.
If you'd like to talk with Aviad while he is here on the 23rd, please
send mail to Dave McKeown@a.
------------------------------
End of AIList Digest
********************
