Psychology in Neural Networks – In Honor of Professor Tracy Mott
Date
2022-09
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
now publishers
Abstract
This paper introduces psychology into neural networks by building a correspondence between the theory of behavioral economics and the theory of artificial neural networks. The connection between these two disparate branches of knowledge is concretely constructed by designing a dictionary between prospect theory and artificial neural networks. More specifically, the activation functions in neural networks can be converted to a probability weighting functions in prospect theory and vice versa. This approach leads to infinitely many activation functions and allows for their psychological interpretation in terms of risk seeking and risk averse behavior.
Description
This is the Accepted Manuscript of the following article: Harpreet Singh Bedi (2022), "Psychology in Neural Networks – In Honor of Professor Tracy Mott", Review of Behavioral Economics: Vol. 9: No. 3, pp 251-262, which has been published in final form at http://dx.doi.org/10.1561/105.00000158. This article is protected by copyright.
Type
Journal Article
item.page.format
Keywords
Citation
Harpreet Singh Bedi (2022), "Psychology in Neural Networks – In Honor of Professor Tracy Mott", Review of Behavioral Economics: Vol. 9: No. 3, pp 251-262. http://dx.doi.org/10.1561/105.00000158