PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

A Statistical Learning Theory Approach of Bloat
Olivier Teytaud, Sylvain Gelly, Nicolas Bredeche and Marc Schoenauer
In: Genetic and Evolutionary Computation Conference, 25 june 2005, Washington D.C. USA.


Code bloat, the excessive increase of code size, is an important is- sue in Genetic Programming (GP). This paper proposes a theoreti- cal analysis of code bloat in the framework of symbolic regression in GP, from the viewpoint of Statistical Learning Theory, a well grounded mathematical toolbox for Machine Learning. Two kinds of bloat must be distinguished in that context, depending whether the target function lies in the search space or not. Then, important mathematical results are proved using classical results from Sta- tistical Learning. Namely, the Vapnik-Cervonenkis dimension of programs is computed, and further results from Statistical Learn- ing allow to prove that a parsimonious fitness ensures Universal Consistency (the solution minimizing the empirical error does con- verge to the best possible error when the number of samples goes to infinity). However, it is proved that the standard method consisting in choosing a maximal program size depending on the number of samples might still result in programs of infinitely increasing size whith their accuracy; a more complicated modification of the fit- ness is proposed that theoretically avoids unnecessary bloat while nevertheless preserving the Universal Consistency.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Learning/Statistics & Optimisation
ID Code:1745
Deposited By:Marc Schoenauer
Deposited On:28 November 2005