Computing as compression
Gerry Wolff
what we are doing at Bangor ...
For the last few years we have been developing ideas about computing
based on the conjecture that:
"All kinds of computing and formal reasoning may usefully be understood
as *information compression* by pattern matching, unification
and search."
A related idea, which has a longer history and is more widely recognised,
is that the storage and processing of information in brains and
nervous systems may often usefully be understood as information compression.
We have also been developing software models of a 'new generation'
computing system dedicated to information compression by
pattern matching, unification and search.
Potential benefits of these developments include:
* Simplification and integration of concepts in computing.
* The seamless integration of different kinds of computing in one
system. 'Kinds of computing' include: learning, information retrieval,
deductive and probabilistic reasoning, execution of functions,
problem solving.
* More flexibility and 'intelligence' in computers.
INFORMATION
A summary of the main ideas is given in the file 'sp_summary' which
may be obtained by anonymous ftp from ftp.sees.bangor.ac, directory
/pub/gerry. Use 'anonymous' as your username and give your email
address (including your normal username) as the password.
The following articles (and one book) give fuller information.
Sources:
* "Computing as compression: overview of the SP theory and system."
Submitted for publication. (This is what it says - an overview or
summary of the SP programme of research as it stands now).
* "Computing as compression: SP20." Submitted for publication.
(Describes a recent model of the SP system attempting to
integrate learning, information retrieval, deductive and
probabilistic reasoning, and other kinds of computing).
* "Towards a new concept of software." Software Engineering Journal
9(1), 27-38, 1994. (Discusses the propositions that the organisation
of 'well structured' software and the execution of software may
both be understood in terms of information compression. Examples are
given from the SP6 model).
* "Computing and information compression: a reply." To appear in
AI Communications, September 1994. (Responds to points made in an
earlier critique, and discusses associated issues).
* "Computing, cognition and information compression." AI Communications
6(2), 107-127, 1993. (Describes the background and motivation for
the research including long-established ideas about economical coding
by brains and nervous systems).
* "Towards a Theory of Cognition and Computing." Chichester:
Ellis Horwood, 1991. (Describes the programme of research up to 1990
including earlier articles on information compression as a mechanism
for language learning).
Gerry Wolff
** Dr J G Wolff, School of Electronic Engineering and Computer Systems,
** University of Wales, Dean Street, Bangor, Gwynedd, LL57 1UT, UK.
** Tel: +44 248 382691. Email: ge...@sees.bangor.ac.uk. Fax: +44 248 361429.
Benjamin R. Zaborowsky
on findings, problems, questions, etc? Are there others like myself who
are the sole support person at your company for neural network implemetation
and who feel the need to bounce ideas back and forth a couple of times
before putting them on the net for the world to see?
If so, I would be happy to communicate with you.
--Ben
-- Ben Zaborowsky be...@ncsa.uiuc.edu or be...@lilly.com
Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285
(317) 276-9918 Fax (317) 277-0786