Miscellaneous Articles Relating to the
SP Framework
Causality, statistical learning and multiple alignment
Paper presented at the UNICOM Tutorial and Seminar on "Causal Models and
Statistical Learning", London, March 1997.
Postscript, Compressed Postscript.
Discusses how "information compression by multiple alignment, unification
and search" may illuminate the learning of causal connections and
the drawing of causal inferences. Describes how the probabilities of
inferences may be calculated.
Learning and reasoning as
information compression by multiple alignment, unification and search
In Computational Learning and Probabilistic Reasoning,
A. Gammerman (Ed.), Chichester: Wiley, 1996, pp 67-85.
An earlier version was presented at Applied Decision
Technologies 95, Brunel University, April 1995.
Postscript, Compressed Postscript.
Discusses how learning and reasoning may be integrated within the
'multiple alignment' framework.
Multiple alignment, cognition and computing
SEECS Report, September 1996.
Postscript, Compressed Postscript,
HTML
(some diagrams may be too wide for printing).
A summary of the multiple alignment proposals.
Computing as compression by multiple alignment, unification
and search (1 & 2)
SEECS Report, July 1995.
Article 1: Postscript, Compressed Postscript;
Article 2: Postscript, Compressed Postscript.
These two articles describe the possible applications of these ideas to pattern recognition,
information retrieval, the execution of simple and composite functions,
forms of deductive and abductive reasoning including inheritance of
attributes and chains of inference, function induction and other
forms of unsupervised learning, and simple forms of problem solving.
|