Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 155
AIList Digest Monday, 23 Jun 1986 Volume 4 : Issue 155
Today's Topics:
Queries - AI Tools Survey & Financial Expert Systems Survey &
Recognition Software & HYDRO & ES Shell & Stereo Vision,
Expert Systems - Validation and Verification,
Resources - Common Lisp Discussion List
Philosophy - Metaphilosophy and Computer Ethics,
Psychology - Doing AI Backwards
----------------------------------------------------------------------
Date: Tue, 17 Jun 86 19:23:19 PDT
From: heher%ford-scf1.arpa@ford-scf1.arpa (Dennis Heher)
Subject: Request for AI Tools Survey
I heard that there is an unclassified report
available that compares all of the commercial
AI tools (KEE, Knowledge Craft, ART, etc.).
This report was supposed to have been generated
at/for Wright-Patterson Air Force Base.
Does anyone have any information (title, report
number, where I can obtain a copy) on such a
report?
Thanks,
Dennis Heher
heher@ford-scf1.arpa
Ford Aerospace & communications Corporation
1260 Crossman Avenue
Sunnyvale, California 94089
(408) 743-3944
------------------------------
Date: 16 Jun 86 16:30:00 GMT
From: pur-ee!uiucdcs!convex!ti-csl!dbdavis@ucbvax.berkeley.edu
Subject: Financial Expert systems survey
I'm looking for a list of systems/software houses that are active
in the development and/or marketing of financial expert systems.
I'd also be curious to know what ( if anything ) the major insurance
companies are up to in terms of in-house development of expert
systems - which companies, and what applications ( risk assessment,
etc. ).
Any help is greatly appreciated. The info will be used as part of a
market survey I'm doing for a class.
--db davis
------------------------------
Date: 19 Jun 86 15:18:21 GMT
From: ihnp4!iwvae!gph@ucbvax.berkeley.edu (haberl)
Subject: AI routines
This is my first time posting to the net. I am doing some research on
Artifical Intellegence processes (Voice Recognition, Text Recognition and
Hand Writing Recognition). If any of you AI wizards can provide me a
reference, on where I can find some routine to provide these services it
would be much appreciated. The references I need are either Information
on Public Domain software or Information on companies that sell software
like this.
THANXS.......
Gregory P. Haberl (312) 979-7072 or (303) 691-4993
Technocrats, Inc.
Po Box 2238 Don't Yet Pathway for return
Littleton, Co 80161
------------------------------
Date: Thu, 19 Jun 86 10:33:30 bst
From: Gordon Joly <gcj%qmc-ori.uucp@Cs.Ucl.AC.UK>
Subject: HYDRO
Does anyone have any information on the HYDRO system,
a water resources management expert system?
Many thanks in advance,
Gordon
INET: gcj%maths.qmc.ac.uk%cs.qmc.ac.uk@cs.ucl.ac.uk
EARN: gcj%UK.AC.QMC.MATHS%UK.AC.QMC.CS@AC.UK
UUCP: ...!seismo!ukc!qmc-ori!gcj
[I'll send Gordon the reference to John Reiter's SRI work
that was included in AIList V4 #141, June 18. -- KIL]
------------------------------
Date: Fri, 20 Jun 86 07:20 CDT
From: Araman@HI-MULTICS.ARPA
Subject: Expert System shell needed for thesis:
I'm doing a Master's thesis concerning combining an expert system shell
and a DBMS. To do this, I need to work with modifying some shell. If
anyone out there has a small, frame or rule based expert system
shell written in LISP or C, and you're willing to give away a copy in
the name of furthering science, send a message to: ARAMAN -at
HI-MULTICS.ARPA Thanks a lot! Sam Levine
------------------------------
Date: Fri, 20 Jun 86 08:11:53 mdt
From: crs%f@LANL.ARPA (Charlie Sorsby)
Subject: Vision Request
I am getting started on a Master's Thesis in the general
area of Computer Vision. Since [...] vision and AI appear
to overlap considerably, I'm trying AIList. I apologize if
this is not an appropriate medium for the following request.
I would sincerely appreciate any pointers to literature and
current research in this area and particularly in the area
of stereo vision. I have Computer Vision by Ballard &
Brown, the vision section of the Handbook of AI, Image
Understanding 1984, Ullman & Richards, eds. and a few
papers that I've found.
What are the current hot research areas in this field?
What, in your opinion, are the most important problems to be
solved? What, aside from stereopsis and range-finding are
options for depth-information recovery? What, currently,
appears to be the best method of obtaining this information
fast enough to be useful?
Is any research being directed at the possibility of real-
time stereo vision? What are your opinions of its feasibil-
ity? Its value?
If any of you have papers that you would be willing to
share, my mailing address is:
Charlie Sorsby
Los Alamos National Laboratory
Post Office Box 1663 MS-J957
Los Alamos, NM 87545
Opinions are welcome and please also mention if I may quote
you or if you prefer that I don't. I would also welcome
suggestions for other lists where it may make sense to make
this request. [Vision-List@ADS is known. --KIL]
While I try to follow the network as time permits, I would
appreciate it if you could mail information to me by way of
one of the paths in my signature.
I will happily summarize any information that I receive and
post it to AIList.
Charlie Sorsby
...{cmcl2, ihnp4, ..}!lanl!crs
crs@lanl.arpa
[I generally forward vision items to Vision-List (and did this time),
but am permitting this message as a favor to Charlie. AI-related
discussion of vision (e.g., for autonomous navigation) is pertinent
to this list, but discussion of particular algorithms would generally
not be. -- KIL]
------------------------------
Date: Fri, 20 Jun 86 14:17 PDT
From: Tom Garvey <GARVEY@SRI-AI.ARPA>
Subject: Re: Expert System Validation and Verification
I think the notion of V&V for expert systems highlights a number
of points about the field. First, in the words of David Mizell
(formerly of ONR), "AI is being overbought." People that should know
better are taking an attitude that there are sufficient useful AI
systems out there that we should be concerned with formal notions of
their capabilities. In point of fact, AI is very much a research topic
(I almost said science), and for most problems we are struggling to find
any solution at all, much less one that will be operationally useful and
verifiable.
The traditional rationale for attempting an "AI" solution to a
problem is that we don't know how to solve the problem directly (if we
did, why screw around), or that our problems come from a large class of
ill-specified problems where flexibility in the problem-solving approach
is of paramount importance (otherwise, ...). AI approaches typically
involve non-deterministic processes such as context-sensitive search
(frequently in large, ill-structured knowledge-bases), and their
performance is therefore extremely difficult to describe much less
quantify. (We don't do a very good job of V&V on deterministic systems
yet.)
Even statistical validation (i.e., try a million random test
cases and measure resulting performance) will be questionable, as
characterizing an appropriate set of test cases spanning the range of
possible or likely inputs will be extremely difficult.
At this point, I view most expert system development as not much
more than programming in a new language. The language offers ease of
specification and representation of certain types of information (oops,
knowledge), but does not lend itself well to either V&V, maintenance, or
robust operation. To the extent that we use expert system developments
to help understand and structure problems, these shortcomings are not
too significant; to the extent that we view the systems as the
solutions themselves, the shortcomings are significant.
All this doesn't help your quest much, but perhaps it will help
lower your expectations.
Cheers,
Tom
------------------------------
Date: Sun, 22 Jun 86 15:05 EDT
From: Brad Miller <miller@UR-ACORN.ARPA>
Subject: Lisp Discussion List
Unfortunately there are few relevant discussion lists on
the Arpanet side of the gateway. We do have one on workstations
and others on particular micros or Lisps, but nothing of the
required generality. ... -- KIL
[...]
Note that there IS a common-lisp mailing list <common-lisp@su-ai.arpa>,
though it is for language definition purposes.
Brad Miller
University of Rochester
miller@rochester.arpa
------------------------------
Date: Mon, 16 Jun 86 16:06:29 EDT
From: "Col. G. L. Sicherman" <colonel%buffalo.csnet@CSNET-RELAY.ARPA>
Subject: Re: Computer Ethics (from Risks Digest)
I have a few comments on _Metaphilosophy,_ as summarized by Bruce Sesnovich:
> The introductory article, James H. Moor's "What is Computer Ethics," is
> an ambitious attempt to define Computer Ethics, and to explain its
> importance. According to Moor, the development and proliferation of
> computers can rightly be termed "revolutionary": "The revolutionary
> feature of computers is their logical malleability. Logical
> malleability assures the enormous application of computer technology."
> Moor goes on to assert that the Computer Revolution, like the
> Industrial Revolution, will transform "many of our human activities and
> social institutions," and will "leave us with policy and conceptual
> vacuums about how to use computer technology."
"Logical malleability" sounds vague to me. If it's just an abstract
phrase for programmability, then I think Moor neglects the real signi-
ficance of computers: that (unlike machines) they accept differing input,
and produce differing output.
I agree fully that computers will cause revolutions. But this talk of
"conceptual vacuums" is born of unavoidable myopia. None of our present-
day prognosticators have shown any serious understanding of the future,
except a few science-fiction writers whom nobody takes seriously. I
suggest that posterity will regard _us_ as the "vacuum" generation,
of an age "when nobody knew how to use computer technology."
> An important danger inherent in computers is what Moor calls "the
> invisibility factor." In his own words: "One may be quite
> knowledgeable about the inputs and outputs of a computer and only dimly
> aware of the internal processing." These hidden internal operations can
> be intentionally employed for unethical purposes; what Moor calls
> "Invisible abuse," or can contain "Invisible programming values":
> value judgments of the programmer that reside, insidious and unseen, in
> the program.
Here Moor appears to be about 30 years behind McLuhan. Try this: "One may
be quite knowledgeable about reading and writing and only dimly aware of
the details of book production and distribution." Or this: "One may be
quite knowledgeable about watching TV and only dimly aware of the physics
of broadcasting." Isn't it rather naive to think that the hidden values
of the computer medium lie in if-tests and do-loops?
To quote one of McLuhan's defocussed analogies: "You must talk to the
medium, not to the programmer. To talk to the programmer is like
complaining to the hot-dog vendor about how badly your team is playing."
Col. G. L. Sicherman
UU: ...{rocksvax|decvax}!sunybcs!colonel
CS: colonel@buffalo-cs
BI: csdsicher@sunyabva
------------------------------
Date: 15 Jun 86 09:08:28 GMT
From: ernie.Berkeley.EDU!tedrick@ucbvax.berkeley.edu (Tom Tedrick)
Subject: Doing AI Backwards (continued)
More on "Doing AI Backwards"
(I can't bear to do anything in a normal way :-)
The exact, concrete nature of models of computation allow
a certain clarity to exist, which was not easily experienced
previously.
Hence, when these models apply to other fields, we may find
a new clarity that was previously lacking.
For example, studying computational complexity has made it
clear that memory can be an expensive resource, and efficient
use of memory of great importance.
Now, can we use this insight to better understand certain phenomena
outside the field of computational devices?
I suggest that memory is also a scarce resource when we take the
human mind as our object of study.
Example: Suppose I am asked to pick up a carton of milk at the
grocery store after work. For some reason this kind
of request irritated me for years, yet I could not
quite pin down the reason for my irritation. I did
not mind walking to the store, spending the money, etc.
It turns out that what bothers me is the use of my
memory to store the request. Thus for the rest of
the day I have less space in my short term memory
for thinking about research, etc. All my work was
made less productive by this misuse of space in
memory.
Hence the individual who asks such seemingly small
favors may be really imposing a heavy cost on his victim.
From a catastrophe theory point of view, we might also
suggest the danger of less efficient thinking due to
reduced space available in memory being magnified into
some larger catastrophe.
Another thing that is clear from studying computational complexity
is that certain problems take more computing time than others.
What insight can we gain about the behavior of the human mind
from this simple idea?
Well, suppose someone asks you a question, expecting a simple
yes or no answer. (Supposedly the truth is simple, so why should
you need to think about the question?)
But suppose you have greater insight into subtle problems posed by
the question than the questioner does. But you need time to
think about it. (By knowing about computational complexity, you
wisely realize that your brain needs to use a few cycles to
figure out what to say.)
Some possibilities:
(1). You answer immediately anyway, yes or no. Then later
one of the subtleties may come back to haunt you, as
the (dumb) questioner comes back to you saying "Well
you said yes, now you are trying to squirm out of it,
you no good scum." Or, "You were not honest with me,
you devious jerk", when you are unable to live up to
your word.
(2). You think for awhile. Then the questioner may think
"Boy is this guy dumb. Can't even answer a simple
question." Or, "This guy is trying to come up with
some kind of a line so as to pull a fast one on me."
Or he may say, "ANSWER THE QUESTION! YES OR NO!"
if he is on a power trip, like, say, a Senate Investigator.
In any case, asking questions and expecting an immediate response,
saying "If you were honest you would not hesitate to answer" is
clearly unfair.
OK, now you can start flaming. But please, for a change, attack what
I said instead of sending hate mail attacking me as an individual.
I have no interest in receiving hate mail.
------------------------------
End of AIList Digest
********************
