Copy Link
Add to Bookmark
Report
AIList Digest Volume 6 Issue 035
AIList Digest Thursday, 18 Feb 1988 Volume 6 : Issue 35
Today's Topics:
References - Genetic Algorithms & Self-Organizing Systems,
AI Tools - Fuzzy Logic vs. Probability Theory,
Nanotechnology - Altering Individual Atoms,
Biography - Richard Greenblat
----------------------------------------------------------------------
Date: 15 Feb 88 18:22:11 GMT
From: g451252772ea@deneb.ucdavis.edu (0040;0000001899;0;327;142;)
Subject: Re: becoming literate with genetic algorithms
The references, generally at good libraries, that I know of for GAs:
Introductory:
Holland, J., et al. INDUCTION. 1986, MIT Press. The book is a coherent
whole, not a collection of separately authored papers - and reads very well
by any standard. Most of it discusses human induction, but the main model
introduced early on is Holland's. And the human material is fascinating in
its own right, only partly because of the lucid presentation. The description
of Holland's GA is complete, and an alternative system, PI, is also presented.
This is a more familiar symbol-based production system, in LISP.
Holland, J. "Genetic Algorithms and Adaptation", in O. Selfridge, et al,
ADAPTIVE CONTROL OF ILL-DEFINED SYSTEMS. 1984, Plenum Press, NY. This is
a discrete chapter, in which an overview of GA is provided. Almost every main
theme is touched on.
Davis, L. GENETIC ALGORITHMS AND SIMULATED ANNEALING. 1987, Morgan Kauffman
Pub, Los Altos, CA. A collection of research papers by Holland's colleagues,
mostly (his INDUCTION chapters are reproduced here also). A good variety of
current work, and again very lucid as technical/research writing goes (by
contrast, the Neural net literature is hopeless). Topics include a study of
the TSP; parallel implementation of the CFS-C simulation library for GA on
the Connection Machine (nice!); Axlerod's study of GA in round-robins of the
iterated Prisoner's dilemma; a somewhat vague but very suggestive study on
designing a mapping from 'an East Asian language' onto a usable keyboard,
using a GA; some formal tests of 'hard' problems for GAs; and another
suggestive paper (for me) on producing long action sequences with GA by
means of 'hierarchical credit allocation' (this problem has parallels in
the animal-behavior literature I'm familiar with).
Holland, J. ADAPTION IN NATURAL AND ARTIFICIAL SYSTEMS. 1975, U. Michigan
Press. The definitive foundation, marred only by a generous use of formal
notation (not insensibly, but offputting nonetheless). The main conceptual
addition since this has been the interpretive change in INDUCTION, I think.
The GA community has held two conferences, last summer and in '86. The
proceedings are available from Lawrence Erlbaum Assoc., 365 Broadway,
Hillsdale, NJ 07642. My copy is on order ("Proc. Second International
Conf. on GA and their applications", held at Cambridge, MA, July 28-81, 1987).
And the various dissertations Holland has supervised are worth perusing via
U.Microfilm copies at $25 each.
For relating GA to NNets, I'll hazard to volunteer Richard Belew's name. He
responded to an earlier posting I made and stated an interest in what
commonalities there might be. He teaches at UCSD: rik@sdcsvax.ucsd.edu.
Oh yes: as the _very best_ intro article to GA, I recommend the final issue
of Science 86, for July, I think. Too bad that mag died.
Hopefully helpfully (let me know what else you find- I've been teaching this
material to budding animal behaviorists!) -
Ron Goldthwaite / UC Davis, Psychology and Animal Behavior
'Economics is a branch of ethics, pretending to be a science;
ethology is a science, pretending relevance to ethics.'
------------------------------
Date: 14 Feb 88 15:31:54 GMT
From: trwrb!aero!venera.isi.edu!smoliar@ucbvax.Berkeley.EDU (Stephen
Smoliar)
Subject: Re: becoming literate with genetic algorithms
In article <9430@shemp.CS.UCLA.EDU> jason@CS.UCLA.EDU () writes:
>John Holland was here recently giving talks on genetic algorithms. I found
>the
>concept rather intriguing. After hearing his lectures, I realized I needed to
>do some introductory reading on the subject to fully appreciate its potential.
>
The best source would be the book entitled INDUCTION, which Holland wrote with
Holyoak, Nisbett, and Thagard. Most of the material from the talk is in
Section 4.1 (I think). The preceding material leading up to the major
argument is very well written, as is the subsequent discussion.
------------------------------
Date: 16 Feb 88 11:16:27 GMT
From: Gilbert Cockton <mcvax!hci.hw.ac.uk!gilbert@uunet.uu.net>
Reply-to: Gilbert Cockton <mcvax!hci.hw.ac.uk!gilbert@uunet.uu.net>
Subject: Re: self organizing systems
In article <8801291421.aa28769@ARDEC-AC4.ARDEC.ARPA> pbeck@ARDEC.ARPA
(Peter Beck, LCWSL) writes:
>
>Is this a generally accepted proposition, ie, that complex constituent
>elements can "NOT" form self organizing systems??
Broadly speaking, social theories are often opposed across a
co-operation vs. conflict continuum. Theories in the Marxian tradition
stress conflict as a fundamental dynamic of society. Theories in the
functionalist tradition stress adaptation towards universal ends (e.g.
Talcott-Parsons). Look to Marxian theories (e.g post/neo/vanilla
-structuralism) for evidence of non-self-organisation. Look to
functionalist ones for evidence of dormitory consensus.
NB Rednecks! - 'Marxian' is a scholarly term, 'Marxist' is both a
scholarly and a political term. Marx claimed he wasn't a Marxist! It
is thus safe to follow up these ideas without the risk of brainwashing
yourself into running off to Cuba/Nicaragua :-)
--
Gilbert Cockton, Scottish HCI Centre, Heriot-Watt University, Chambers St.,
Edinburgh, EH1 1HX. JANET: gilbert@uk.ac.hw.hci
ARPA: gilbert%hci.hw.ac.uk@cs.ucl.ac.uk UUCP: ..{backbone}!mcvax!ukc!hci!gilbert
------------------------------
Date: Mon, 15 Feb 88 20:46:33 PST
From: golden@frodo.STANFORD.EDU (Richard Golden)
Subject: FUZZY LOGIC VS. PROBABILITY THEORY
I am not an expert in Fuzzy Logic or Probability Theory but I have examined
the literature regarding the foundations of Probability Theory and the
derivation of these foundations from basic principles of deductive logic.
The basic theoretical result is that selecting a "most probable" conclusion
for a given set of data is the ONLY RATIONAL selection one can make in
an environment characterized by uncertainty. (Rational selection in this
case meaning consistency with the classic deductive/symbolic logic - boolean
algebra.) Thus, one could argue that if one constrains the class of
possible inductive logics to be consistent with the laws of deductive logic
then Probability Theory is the MOST GENERAL type of inductive logic.
The reference from which these arguments are based is given
by Cox (1946). Probability Frequency and reasonable expectation.
American Journal of Statistical Physics, 14, 1-13.
The argument is based upon the following hypotheses:
(i) The belief of the event B given A may be represented by a
real-valued function F(B,A).
(ii) F(~B,A) may be computed from F(B,A)
(iii) F(C and B,A) may be computed from F(C,B and A) and F(B,A)
Note this assumption's similarity to Bayes Rule but the
multiplicative property is not assumed.
(iv) Assumptions (i), (ii), and (iii) must be consistent with the
laws of Boolean Algebra (i.e., deductive/symbolic logic).
>From these assumptions one can prove that F(B,A) must be equivalent
to the conditional probability of B given A. That is, F(B,A) must
lie between a maximum and minimum value (say 1 and 0) and the sum of
all possible values for B for a particular value of A must equal the
maximum value (1). Note that we are taking the subjectivist view of
probability theory and we are NOT interpreting the probability of
an event as the limiting value of the relative frequency of an event.
To my knowledge, the axioms of Fuzzy Logic can not be derived from
consistency conditions generated from the deductive logic so I conclude
that Fuzzy Logic is not appropriate for inferencing. Any comments?!!!
Richard Golden
Psychology Department
Stanford University
Stanford, CA 94305 GOLDEN@PSYCH.STANFORD.EDU
Cc:
------------------------------
Date: Mon, 15 Feb 1988 18:31 EST
From: MINSKY%OZ.AI.MIT.EDU@XX.LCS.MIT.EDU
Subject: AIList V6 #27 - Nanotechnology
Dolata's reamrks about nanoscopic chemistry missed the point, so far
as I can see, in arguing that because it is a scanning microscope it
is not involved with individual molecules but is more like regular
volume chemistry. However, the molecular rearrangement was not
accomplished by a conventional bulk effect. Instead, it was
accomplished by a sub-microsecond pulse applied during the scan so
that it occurred while the needle was over a particular molecule. The
next step, of course, is to try to make a particular modification at a
particular site on the molecule.
Much more will be done in the area soon, I'm sure, because the
techniques seem quite accessible. But I see no reason to denigrate
the technique because it uses scanning. Simply think of scanning as
examining, and possibly modifying, large numbers of points is
sequence. What could be better? The trouble with traditional
chemistry is, in fact, that it is constrained to do the same thing to
everything, in parallel.
------------------------------
Date: 14 Feb 88 08:28
From: minow%thundr.DEC@decwrl.dec.com (Martin Minow THUNDR::MINOW
ML3-5/U26 223-9922)
Subject: Article on Richard Greenblat
AIList readers might enjoy this article from the Boston Globe, Feb. 7, 1988.
ZORCHED OUT: A COMPUTER HACKER'S TALE
by Alex Beam, Boston Globe staff
Richard Greenblatt: Single-minded, unkempt, prolific, and
canonical MIT hacker who went into night phase so often
that he zorched his academic career. The hacker's hacker.
- HACKERS by Steven Levy.
CAMBRIDGE -- "Lights On!" Greenblatt yells, pushing through the
door of MIT's Model Railroad Club. "That's just in case
anybody's sleeping under the layout." He explains to a visitor.
"They might pick up a shock or something."
Happily, no one is sleeping underneath the thousand feet of
handmade track that may be the world's most sophisticated model
railway. The last person to fall asleep under the layout was
probably Greenblatt, who spent so much tinkering - "hacking" -
with the railroad's switching system, and with his other
favorite toy, computers, that he flunked out of MIT in his
sophomore year.
Greenblatt, now 44, has gone on to bigger things. After a long
career as senior researcher at MIT's Artificial Intelligence Lab,
he helped found Lisp Machine Inc., one of the first artificial
intelligence startups. Now he is president of Cambridge-based
GigaMOS, which purchased LMI's assets after it went broke last
year.
But scratching the surface of Richard Greenblatt, AI entrepreneur,
one quickly finds traces of 17-year-old Ricky Greenblatt, the
soda-pop swilling science whiz who arrived at MIT as a bewildered
freshman from Columbia, MO, in 1963.
Greenblatt still drops in on the railroad club from time to time,
and exudes boyish enthusiasm when demonstrating "the famous
Greenblatt track cleaning machine," a cleverly-engineered
locomotive that spins an abrasive grinding wheel over the
nickel-silver track.
He sheepishly explains that he is "out of phase" on a particular
day, because he spent the previous night hacking away on a
computer.
And even though he has cleaned up his presentation - friends say
he bathed so rarely as an undergraduate that they had to ambush
him with air freshener - Greenblatt still acts like an
absent-minded computer genius. Pallid-skinned from long hours of
computer work, he trundles around Cambridge in rumpled work
pants and a plaid shirt, with a digital calculator watch
protruding from his breast pocket and a cellular phone slung
across his shoulder.
Although he has earned plenty of money in his computer ventures,
he still rents a room in the same house in Belmont where he has
lived for 20 years. Why not buy a house: "It's too much
trouble," Greenblatt says. "You have to pay taxes, mow the lawn.
I don't want to bother."
"Ricky lives in a world of his own, dominated by his own genius,"
says Andy Miller, who briefly roomed with Greenblatt at MIT. "We
never saw him when he lived with us. The Sun meant absolutely
nothing to him - it happened to rise and fall in a way that
wasn't in synch with his schedule."
After two semesters on the Dean's List at MIT, Greenblatt threw
in with the small band of electronics fanatics hanging around
the Model Railroad Club. Synchronizing the model railroad's
switching system - its circuits can control five trains chugging
across the vast layout, and set the 200 switches so no crashes
occur - turned out to be a lot like programming the early
computers that were making their first appearance in MIT labs.
(It also resembled another electronic gimmick called "phone
hacking," or fooling the phone system into placing free
long-distance calls, which resulted in suspension of several of
Greenblatt's friends.)
Greenblatt and his friends often spent the daylight hours working
on the railroad, and then migrated to a neighboring lab to stay
up all night next to the PDP-1, DIGITAL EQUIPMENT CORP.'s first
computer. Fueled by the Railroad Club's private Coca-Cola
machine, Greenblatt and his fellow hackers "wrapped around" day
into night, working for 30 hours at a stretch to solve thorney
problems, either with the railway or the computer.
"To a large extent, our group wasn't interested in the normal
social events around the institute," explains fellow hacker Alan
Kotok, now a corporate consulting engineer at Digital. "The
railroad club was like a fraternity. There were people you could
talk to day or night about things of common interest.
Although no one asked him to, Greenblatt wrote a high-level
language computer program for the PDP-1, so the club's timetable
system could be stored on the new computer. Unfortunately, the
young programmer's deepening involvement in computer hacking
doomed his academic career. "I sort of zorched out on classes,"
Greenblatt admits. During one of his 30-hour work blasts,
Greenblatt slept through a final exam, and had to leave MIT.
Of course, MIT didn't get where it is today by turning away
computer talent. After a brief sojourn on Route 128, Greenblatt
landed a job as a programmer at the Artificial Intelligence Lab,
and stayed for 20 years.
Greenblatt's fame grew and grew. He and a co-worker wrote ITS,
and early minicomputer time-sharing program that is still in use
today. He was one of the early programmers to work in LISP, the
high-level language that has become the key building block for
artificial intelligence.
"He would attack problems with great vigor," remembers Donald
Eastlake, another railroad club alumnus. "Everybody was smart,
but the people who really excelled were smart and tenacious.
He was one of the primary examples of that."
An accomplished chess player, Greenblatt wrote MacHack, a chess
program for a later DIGITAL mini, the PDP-6. The program scored
an important victory for AI boosters when it defeated a
prominent critic of artificial intelligence who insisted that a
computer would never play chess well enough to beat a 10-year
old. The program later became a member of the American Chess
Federation and the Massachusetts State Chess Association.
When Greenblatt later did graduate work at MIT, administrators
hinted that if he submitted his chess program as a doctoral
thesis, he might be awarded a degree. "I never really got around
to it," Greenblatt confesses. "It just didn't seem that important."
------------------------------
End of AIList Digest
********************
