Copy Link
Add to Bookmark
Report
Neuron Digest Volume 02 Number 10
NEURON Digest Tue Apr 14 12:25:43 CDT 1987 Volume 2 - Issue 10
Today's Topics:
diff "connectionist computing" "distributed computing"
Acquisition of Baseline Neurocomputer
Philosophy/Psychology Meeting
NLP: A PDP Approach to Discovering the Hidden Structure of Speech
Seminar - A Connectionist Representation for Concept Structures (Rutgers)
----------------------------------------------------------------------
Date: 3-APR-1987 14:46
From: @C.CS.CMU.EDU:JOSE@LATOUR.ARPA
Subject: abstract
Knowledge Representation in Connectionist Networks
Stephen Jose Hanson and David J. Burr
Bell Communications Research
Morristown, New Jersey 07960
Abstract
Much of the recent activity in connectionist models stems
from two important innovations. First, a layer of
independent, modifiable units (hidden layer) that can model
the statistics of the domain and in turn perform significant
associative mapping between stimulus pairs. Second, a
learning rule that dynamically creates representation in the
hidden layer based upon constraints from a teacher
signal. Both Boltzmann machine and back-propagation models
share these two innovations and interestingly ones that
were apparently well known by Rosenblatt[14]. Although
presently, many complex perceptual and cognitive models
have been constructed using these methods the exact
computational nature of the networks in terms of their
clustering, partitioning, and generalization behavior is not
well understood.
In this paper we present a uniform view of the
computational power of multi-layered learning (MLL)
models. We show that MLL models represent knowledge by
applying Boolean combination rules to partition the problem
space into regions. A by-product of these rules is that
knowledge is represented as distributed patterns of
activation in the hidden layers. Their partitioning
capability is related to both the neural device model and
the network complexity in terms of numbers and layers of
neurons. The device model determines the shape of an
elementary boundary segment and the network determines how
to combine the segments into region boundaries.
For continuous problem spaces two hidden layers are
sufficient to form arbitrary regions (or Boolean functions)
in the space, and for binary-valued spaces a single layer
suffices. Finally we show that networks can produce
probabilistic combination rules which closely approximate
the Bayes risk.
You can get a copy of this paper by replying to this message
or writing to jose@bellcore or djb@bellcore, comments
appreciated.
------------------------------
Date: 16 Mar 87 12:40:15 GMT
From: Dekang Lindek <mcvax!cs.strath.ac.uk!lindek@seismo.CSS.GOV>
Reply-to: lindek@cs.strath.ac.uk (Dekang Lindek)
Subject: diff "connectionist computing" "distributed computing"
Any one know the result of the title of this article?
advThanksance.
!-@-#-$-%-^-&-*-(-)-!-@-#-$-%-^-&-*-(-)-!-@-#-$-%-^-&-*-(-)
Dekang Lin
Dept. of CS
Univ. of Strathclyde
26 Richmond St.
Glasgow, G1 1XH, U.K.
lindek%cs.strath.ac.uk@ucl-cs.arpa
....!seismo!mcvax!ukc!strath-cs!lindek
------------------------------
Date: Mon 13 Apr 87 11:52:02-PDT
From: Ken Laws <Laws@stripe.sri.com>
Subject: Humor
Subject: Acquisition of Baseline Neurocomputer
Thought you'd enjoy seeing this birth announcement I sent around SRI.
-- Ken
This is to announce the availability of a conventional-style
neural-network cytocomputer for training and observation. The
project was delivered late Saturday after months of intensive
development. The prototype, known as a "Devon Lee", weighs
about 6.3 pounds and has a maximum dimension of 19 inches; this
may grow as capabilities are added. Communication is currently
a bottleneck, but we expect to begin teaching it chess endgames
as soon as the perceptual networks are trained. Considerable
credit is due to my coworker, Lily Laws, for concluding this
acquisition two and a half weeks ahead of schedule and with no
expenditure of IR&D funds.
-- Ken
------------------------------
Date: 13-APR-1987 15:46
From: SIMPSONP@COD.NOSC.MIL
Subject: ANS Survey Paper
A Survey of Artificial Neural Systems
Patrick K. Simpson
Unisys
San Diego Systems Engineering Center
4455 Morena Boulevard
San Diego, CA 92117
619/483-0900
Abstract
This paper is a survey of the field of Artificial
Neural Systems (ANSs). ANSs have a large number of highly
interconnected processing elements that demonstrate the
ability to learn and generalize from presented patterns.
ANSs represent a possible solution to previously difficult
problems in areas such as speech processing and natural
language understanding. This paper presents a brief history
of ANSs, examples of ANS models and areas where the technol-
ogy has been applied. Also discussed is the connection
between Artificial Intelligence (AI) and ANS, computer
architectures that are evolving from this field, and two ANS
algorithms.
[Copies are available from simpson@cod.nosc.mil or the address
listed above - MTG]
------------------------------
Date: 3-APR-1987 07:42
From: MCDONLEY%CODSD1@TI-CSL.CSNET
Subject: TRW MARK III-1 ANS processor
I just received a Product Description:
THE TRW MARK III-1
ARTIFICIAL NEURAL SYSTEM PROCESSOR
6 JAN 87
619-592-3381 (SAN DIEGO)
This is a coprocessor for vax family machines with color display and a
development environment. Benchmark claim (no description of bench) is 29
times faster than same bench in Pascal on micro-VAX. Uses 8 bit unsigned
values, 21 bit integer weights, up to 65K "processing elements" and up to
1.1Million interconnects with trainable weights. Claims network update rates
of several times per second "In many typical cases, where full connectivity is
not required".
------------------------------
From: harnad%mind@princeton.csnet
Date: 4 Apr 87 05:43:35 GMT
Subject: Philosophy/Psychology Meeting
Keywords: 13th Annual SPP Meeting, UCSD. June 21 - 23, Program follows:
Program of the 13th Annual Meeting of the Society for Philosophy and Psychology
June 21 -23, University of California, San Diego
For program information: William Bechtel (SPP Program Chairman),
Philosophy Department, Georgia State University, Atlanta GA 30303-3083
phone: (404)-658-2277 bitnet address: psuvax1!phlpwb%GSUMVS1.BITNET
For membership information: Patricia Kitcher, Philosophy Department,
University of California-San Diego, La Jolla CA 92093
arpanet address: sdcsvax!ir205%sdcc6
-------- SUNDAY, JUNE 21, 1987 --------
9:00 - 11:00am SYMPOSIUM: DEPRESSION, COGNITION, AND RATIONALITY
Chair: Evalyn Segal, Psychology, San Diego State University
Speakers: George Graham, Philosophy, University of Alabama at Birmingham
Christopher Peterson, Psychology, University of Michigan
Lynn Rehm, Psychology, University of Houston
Commentator: Richard Garrett, Philosophy, Bentley College
1:00 - 3:15pm CONCURRENT CONTRIBUTED PAPERS SESSIONS I AND II
SESSION I: Behavior and Belief
Chair: James Pate, Psychology, Georgia State Unviersity
Speaker: Ruth Garrett Millikan, Philosophy, University of Connecticut
"What is Behavior? or Why Narrow Psychology/Ethology is Impossible"
Commentator: John Biro, Philosophy, University of Oklahoma
Speaker: David Martel Johnson, Philosophy, York University
"'Brutes Believe Not': Why Non-Human Animals Have No Beliefs"
Commentator: Carolyn Ristau, Psychology, Vassar
SESSION II: Computational Theories of Mind
Chair: Owen Flanagan, Philosophy, Wellesley
Speaker: David Kirsch, Artificial Intelligence, MIT
"The Concept of Computation in Connectionist Systems"
Commentator: Brian Cantwell Smith, Computer Science, Xerox PARC
Speaker: Joseph Levine, Philosophy, North Carolina State University
"Demonstrative Thought"
Commentator: La Verne Shelton, Educational Testing Service, Princeton
3:30-5:00pm INVITED LECTURE: LANGUAGES OF THE DEAF
Chair: Adele Abrahamsen, Language Research Center, Georgia State
Speaker: Howard Poizner, Salk Institute, San Diego
"Brain Function for Language: Perspectives from Another Modality"
7:00-10:00pm SYMPOSIUM: ANALOGY AND LEARNING
Chair: Paul Thagard, Cognitive Science, Princeton
Speakers: Dedre Gentner, Psychology, University of Illinois
Doug Medin, Psychology, University of Illinois
Keith Holyoak, Psychology, University of California, Los Angeles
Commentator: Eva Kittay, Philosophy, SUNY, Stony Brook
----- MONDAY, JUNE 22, 1987 --------
9:00-11:30am SYMPOSIUM: CONNECTIONISM AND IMAGE SCHEMATIC STRUCTURES
Chair: Patricia Churchland, Philosophy, University California, San Diego
Speakers: David Rumelhart, Psychology, University of California, San Diego
George Lakoff, Linguistics, University of California, Berkeley
Mark Johnson, Philosophy, Southern Illinois University
Terrence Sejnowski, Biophysics, Johns Hopkins University
12:30-2:45pm CONCURRENT CONTRIBUTED PAPERS SESSIONS III, IV, AND V
Session III: Logic and Reasoning
Chair: Ralph Kennedy, Philosophy, Wake Forest
Speaker: David Sanford, Philosophy, Duke University
"Circumstantial Validity"
Commentator: John Rust, Psychology, London School of Education
Speaker: Howard Margolis, Committee on Public Policy, University of Chicago
"Habits of Mind"
Commentator: Stuart Silvers, Philosophy, Tilburg University
Session IV: Mentalistic Explanations
Chair:
Speaker: Joseph Thomas Tolliver, Philosophy, University of Maryland
"Knowledge Without Truth"
Commentator: Kent Bach, Philosophy, San Fransciso State University
Speaker: Louise M. Antony, Philosophy, North Carolina State University
"Anomalous Monism and the Problem of Explanatory Force"
Commentator: Ken Presting, Philosophy, San Francisco State University
Session V: Subjective Experience
Chair: Hilary Kornblith, Philosophy, Vermont
Speaker: James S. Kelly, Philosophy, Miami University
"On Quining Qualia"
Commentator: Henry Jacoby, Philosophy, East Carolina University
Speaker: Richard J. Hall, Philosophy, Michigan State University
"Is An Inverted Pain-Pleasure Spectrum Possible?"
Commentator:
3:00-5:30pm SYMPOSIUM: CONCEPTUAL AND SEMANTIC CHANGE IN CHILDHOOD AND SCIENCE
Chair:
Speakers: Annette Karmiloff-Smith, MRC, Cognitive Development Unit
Alison Gopnik, Psychology, University of Toronto
Susan Carey, Psychology, Massachusetts Institute of Technology
Philip Kitcher, Philosophy, University of California, San Diego
8:00-9:00pm PRESIDENTIAL ADDRESS
Chair: Alvin Goldman, Philosophy, Arizona
Speaker: Stevan Harnad, Behavioral and Brain Sciences
"Uncomplemented Categories, or, What Is It Like To Be a Bachelor?"
------ TUESDAY, JUNE 23, 1987 ------
9:00-11:00am SYMPOSIUM: SEMANTICS
Chair: Richard Jeffrey, Philosophy, Princeton
Speakers: Mark Johnston, Philosophy, Princeton
Barbara Hall Partee, Linguistics, U. Massachusetts, Amherst
Norbert Hornstein, Linguistic, University of Maryland
Commentator: Stephen Schiffer, Philosophy, University of Southern California
11:15-12:30pm INVITED LECTURE: Memory and Brain
Chair:
Speaker: Larry R. Squire, Psychiatry, University of California, San Diego
"Memory and Brain: Neural Systems and Behavior"
1:30-3:45pm CONCURRENT CONTRIBUTED PAPER SESSIONS VI AND VII
SESSION VI: CONCEPTS
Chair: Bernard Kobes, Philosophy, Arizona State University
Speaker: Kenneth R. Livingston & Janet Andrews, Psychology, Vassar College
"Reflections on the Relationship Between Philosophy and Psychology in the Study of Concepts?: Is there Madness in our Methods?"
Commentator: Robert McCauley, Philosophy, Emory University
Speaker: Andrew Woodfield, Philosophy, Bristol
"A Two-Tiered Model of Concept Formation"
Commentator:
SESSION VII: INTENTIONALITY
Chair: Douglas G. Winblad, Philosophy, Georgia State University
Speaker: Ron Amundson, Philosophy, University of Hawaii at Hilo
"Doctor Dennett and Doctor Pangloss"
Commentator: Justin Leiber, Philosophy, University of Houston
Speaker: Robert Van Gulick, Philosophy, Syracuse
"Consciousness, Intrinsic Intentionality,
and Self- Understanding Machines"
Commentator: Nick Georgalis, Philosophy, East Carolina University
4:00-5:30pm INVITED LECTURE: CONSCIOUSNESS
Chair:
Speakers: Daniel Dennett, Philosophy, Tufts University
Kathleen Akins, Philosophy, Tufts University
BEACH PARTY
--
Stevan Harnad (609) - 921 7771
{bellcore, psuvax1, seismo, rutgers, packard} !princeton!mind!harnad
harnad%mind@princeton.csnet harnad@princeton.ARPA harnad@mind.Princeton.EDU
------------------------------
Date: 3 April 1987 1734-PST (Friday)
From: gross@nprdc.arpa
Subject: NLP: A PDP Approach to Discovering the Hidden Structure of Speech
SICNATLANG MEETING
(SICNATLANG = Special Interest Committee in Natural Language)
(SIGART is part of ACM and SICNATLANG is part of SIGART!)
Date: April 6
Time: 6:30 p.m.
Place: UCSD, Psych and Ling Bldg., Rm 3118
Speaker: Dr. Jeff Elman, UCSD Linguistics Dept.
Title: "Discovering the Hidden Structure of Speech: A Parallel
Distributed Processing Approach"
As part of his talk, Jeff will play a tape that demonstrates
the gradual, marked optimization of synthesized speech that
occurs as the neural network stabilizes.
See you there! Contact Beth Sundheim or Michelle Gross for further
information.
Email addresses: sundheim@nosc.arpa, gross@nprdc.arpa.
phone: Beth at work on Monday at 225-7778.
Michelle: 225-6436
------------------------------
Date: Mon, 6 Apr 87 17:35 EDT
From: FAWCETT@RED.RUTGERS.EDU
Subject: Seminar - A Connectionist Representation for Concept Structures (Rutgers)
On Tuesday, May 5th at 11 AM, David Touretzky of Carnegie-Mellon University
will present a talk on "A Connectionist Representation for Concept
Structures". The abstract follows. The room will be announced shortly.
--------------------
A Connectionist Representation for Concept Structures
David S. Touretzky
Computer Science Department
Carnegie Mellon University
I will describe some work I am doing with Shai Geva on a representation for
frame-like structures in a neural network called DUCS. Slot names and slot
fillers are diffuse patterns of activity spread over a collection of units.
The choice of a distributed representation gives rise to certain useful
properties not shared by conventional frame systems. One of these is the
ability to encode fine semantic distinctions as subtle variations on the
canonical pattern for a slot.
DUCS typically maintains several concepts simultaneously in its concept memory.
It can associatively retrieve a concept given one or more slots as cues. I
will show how Hinton's notion of a ``reduced description'' can be applied to
make one concept fill a slot in another.
------------------------------
Date: 7 Apr 87 01:54:37 EDT
From: KALANTARI@RED.RUTGERS.EDU
Subject: A Connectionist Approach to Word Sense Disambiguation (Rutgers?)
Computer Science Department Colloquium :
DATE: Friday, April 10, 1987
SPEAKER: Dr. Gary Cottrell
TITLE: A Connectionist Approach to Word Sense Disambiguation
AFFILIATION: UC San Diego / University of Rochester
TIME: 2:50 (Coffee and Cookies will be setup at 2:30)
PLACE: HILL CENTER 705
A model of the parsing of natural language is proposed which conforms
to psycholinguistic, neurolinguistic and computational constraints.
The parsing model uses a particular spreading activation or neural
network scheme called connectionism which entails a massive number of
appropriately connected computing units that communicate through
weighted levels of excitation and inhibition. The model serves as a
demonstration that connectionist models provide a useful architecture
for representing and using the large number of constraints used in
disambiguating words. Psychological data on the access of word
meanings and some recent neurolinguistic data are accounted for. The
model, in turn, has implications for psycholinguistic and
neurolinguistic research.
------------------------------
End of NEURON-Digest
********************
