Complexity International       /vol09/dopazo01/ © Copyright 2002     
Volume 09 Received: 
Accepted: 
7 Aug 2002
11 Oct 2002



Mutual influence of learning and evolution

Dopazo, H.J. & Perazzo, R.P.J.

Abstract
     We review several models that deal with the interaction of the genetic and behavioral systems throughout evolution. We present the conjecture of Baldwin and the subsequent developments of his ideas and those developed by Schmalhausen. We discuss the experiments of genetic assimilation of Waddington and the modern viewpoint based in the increasingly important role of the modification that the individuals do of the environment in which they dwell. Finally we consider the theoretical model of Hinton & Nowlan and an extension that involves a population of adaptive perceptrons in which some of their synapses can be updated through a learning process. This new approach leads to a halting of the Baldwin Effect due to efficient learning. This causes that the transcription of environmental data into genetic information remains hindered by learning, instead of stimulated as it is usually understood. From the perspective of the general description of complex systems, the interaction of learning and evolution can be considered as the mutual influence between two different hierarchical levels of selection taking place in two widely different timescales.


Full Text

Multimedia Links
(none)

Reference Links
  • http://www.ai.mit.edu/~joanna/baldwin.html

    Citation Reference
    Dopazo, H.J. & Perazzo, R.P.J. (2002), Mutual influence of learning and evolution, Complexity International, Volume 09, Paper ID: dopazo01, URL: http://www.complexity.org.au/vol09/dopazo01/
  •      Get viewers
    for PS & PDF

    Aladdin GhostScript

    Adobe Acrobat