Copy Link
Add to Bookmark
Report
NL-KR Digest Volume 01 No. 30
NL-KR Digest (12/22/86 11:28:25) Volume 1 Number 30
Today's Topics:
Responses to Query on Books on Parsing
Seminar - Uncertainty in AI: Is Probability Adequate (MIT)
Seminar - Linde on Structural Semantics
Princeton seminar (past) - Connectionism and Cog. Linguistics
----------------------------------------------------------------------
Date: 17 Dec 86 20:07:08 GMT
From: rutgers!atux01!jlc@lll-crg.arpa (J. Collymore)
Subject: Responses to Query on Books on Parsing
A few weeks ago I posted a request for information on books that talked
about parsing principles in some detail. I got more requests on posting my
responses than I actually got in responses. However, I am hereby posting those
reponses that I did receive.
Thanks again to those of you who responded with information.
Jim Collymore
===============================================================================
>From seismo!unido!ecrcvax!tomasic
Date: Thu, 4 Dec 86 10:18:33 -0100
From: "Anthony Tomasic" <rutgers!seismo!unido!ecrcvax!tomasic>
Subject: natural language
A good book which addresses natural language understanding in a PROLOG
context is:
Natural Language Access to Databases
by Mark Wallace
Ellis Wood Publishing (1982?)
>From akgua!bullwinkle!cornell!belmonte@svax.cs.cornell.edu
Date: Sun, 7 Dec 86 00:55:17 EST
From: ihnp4!bullwinkle!cornell!belmonte@svax.cs.cornell.edu (Matthew Belmonte)
Subject: Re: References needed for learning parsing principles
Organization: Cornell Univ. CS Dept.
_Compiler_Construction:__Theory_and_Practice_
William A. Barrett, Rodney M. Bates, David A. Gustafson, John D. Couch
Science Research Associates, 1986 (2nd edition)
This is what gave me a lot of the background I needed for an internship I had,
wworking on a natural language parsing project that used a CFG to represent
a very restricted, but English-sounding, subset of English.
It does not address any implementations in terms of languages and operating
systems, but it does, I feel, present parsing concepts in a clear way.
It is not, however, specific to the parsing of natural languages.
--
"The spirit is willing but the flesh is under court injunction."
Matthew Belmonte
ARPA: <belmonte@svax.cs.cornell.edu>
BITNET: <d25y@cornella> <d25y@crnlvax5>
UUCP: ..!decvax!duke!duknbsr!mkb
------------------------------
Date: Tue, 16 Dec 86 02:14:58 EST
From: "Steven A. Swernofsky" <SASW%MX.LCS.MIT.EDU@MC.LCS.MIT.EDU>
Subject: Seminar - Uncertainty in AI: Is Probability Adequate (MIT)
[Excerpted from AIList]
UNCERTAINTY IN AI:
IS PROBABILITY EPISTEMOLOGICALLY AND HEURISTICALLY ADEQUATE?
MAX HENRION
Carnegie Mellon
ABSTRACT
New schemes for representing uncertainty continue to
proliferate, and the debate about their relative merits seems to
be heating up. I shall examine several criteria for comparing
probabilistic representations to the alternatives. I shall
argue that criticisms of the epistemological adequacy of
probability have been misplaced. Indeed there are several
important kinds of inference under uncertainty which are
produced naturally from coherent probabilistic schemes, but are
hard or impossible for alternatives. These include combining
dependent evidence, integrating diagnostic and predictive
reasoning, and "explaining away" symptoms. Encoding uncertain
knowledge in predictive or causal form, as in Bayes' Networks,
has important advantages over the currently more popular
diagnostic rules, as used in Mycin-like systems, which confound
knowledge about the domain and about inference methods.
Suggestions that artificial systems should try to simulate human
inference strategies, with all their documented biases and
errors, seem ill-advised. There is increasing evidence that
popular non-probabilistic schemes, including Mycin Certainty
Factors and Fuzzy Set Theory, perform quite poorly under some
circumstances. Even if one accepts the superiority of
probability on epistemological grounds, the question of its
heuristic adequacy remains. Recent work by Judea Pearl and
myself uses stochastic simulation and probabilistic logic for
propagating uncertainties through multiply connected Bayes'
networks. This aims to produce probabilistic schemes that are
both general and computationally tractable.
HOST: PROF. PETER SZOLOVITS
------------------------------
Subject: Seminar - Linde on Structural Semantics
Date: Wed 17 Dec 86 17:46:25-EST
From: LPOLANYI at G.BBN.COM
Speaker: Dr. Charlotte Linde
Structural Semantics
Palo Alto, CA
Topic: Politeness and Accidents:
The Linguistic Study of Aviation Discourse
When: Monday December 22, 1986
3:00 PM
Where: BBN Labs
10 Moulton Street
Cambridge, MA
2nd Floor Large Conference Room
It has been shown that one major cause of accidents in commercial
aviation is poor crew communication and coordination. This talks
reports on the results of several projects attempting to specify
patterns of communication which are associated with unsuccessful
flights.
This study of communicative success uses as its data transcripts of 8
aviation accidents, the "black box" transcripts, as well as
transcripts of 14 crews run through the same problem situation in a
full-mission flight simulator. The linguistic variable considered is
mitigation; the real world variables are success or failure of the
individual communication, and peer judgments of the overall
effectiveness of the simulator crews. Our hypotheses about the
success or failure of individual utterances require the discrimination
of a number of degrees of mitigation. A 4 degree scale of mitigation
was established by using the judgments of several linguistic analysts,
and was validated against the judgments of members of the aviation
community.
Using this scale, a number of hypotheses were confirmed: (1)
Utterances going up the chain of command are more mitigated than those
going down, showing that mitigation is sensitive to social rank. (2)
Mitigation is sensitive to crew members' perception of the seriousness
of the situation. (3) Rate of planning and explanation is sensitve to
crew members' perception of the seriousness of the situation. (4)
Suggestions by a crew member to the captain are more likely to fail if
they are mitigated than if they are direct. (5) Utterances
introducing a new topic are more likely to fail if they are mitigated
than if they are direct.
These findings show that lingusitic methodology is a valuable tool in
understanding problems in an important real-world setting, and suggest
directions for training of aviation crews in effective communication
patterns.
------------------------------
Date: 9 Dec 86 11:38:40 EST
From: Tom Fawcett <FAWCETT@RED.RUTGERS.EDU>
Subject: Princeton seminar (past) - Connectionism and Cog. Linguistics
Resent-Message-Id: <12263423021.12.LAWS@SRI-STRIPE.ARPA>
TITLE: Connectionism and Cognitive Linguistics
SPEAKER: George Lakoff, University of California, Berkeley
DATE: Monday, December 8
LOCATION: Princeton University, Green Hall, Langfeld Lounge
TIME: 12 Noon
In the 1970's Cognitive Science developed largely under the assumption that
human reason could be characterized in terms the manipulation of symbols
that were to get their meaning via a relation to things in the world. This
view grew out of the attempt to use mathematical logic and model theory as a
basis for the study of human reasoning. Over the past decade it has become
increasingly clear that such an approach cannot work. In its place there has
developed a cognitive approach to semantics. This talk will (1) provide an
overview of the phenomena that have led to the development of cognitive
semantics, (2) survey the mechanisms used in cognitive semantics, and (3)
discuss how such mechanisms might be made sense of within the emerging
connectionist theory of mind.
------------------------------
End of NL-KR Digest
*******************
