Artificial life, natural rationality and probability matching

Author

Hardy-Vallée, Benoit

Author_Institution

Dept. of Philos., Waterloo Univ., Ont.

fYear

2007

fDate

1-5 April 2007

Firstpage

123

Lastpage

129

Abstract

Three different agendas for the philosophy of ALife are presented. While the metaphysical and the epistemological ones have been extensively developed, a third one is proposed here. The naturalistic agenda is more interested in resolving empirical problems. Consequently, this paper addresses an empirical issue: the adaptive function of a phenomenon known as "probability matching", that is, the tendency to match the probability of choice with the probability of reward. Probability matching and its common interpretations are presented and discussed. Based on ALife and neural networks simulations, the author presents and alternative hypothesis: probability matching is an adaptation to radical uncertainty

Keywords

adaptive systems; artificial life; neural nets; pattern matching; probability; ALife simulations; adaptive function; artificial life; choice probability; natural rationality; naturalistic agenda; neural networks simulations; probability matching; radical uncertainty; reward probability; Artificial intelligence; Chromium; Gold; Intrusion detection; Light rail systems; Radiofrequency interference; Tellurium; Tin;

fLanguage

English

Publisher

ieee

Conference_Titel

Artificial Life, 2007. ALIFE '07. IEEE Symposium on

Conference_Location

Honolulu, HI

Print_ISBN

1-4244-0701-X

Type

conf

DOI

10.1109/ALIFE.2007.367787

Filename

4218877