Spike-based decision learning of nash equilibria in two-player games

Friedrich, Johannes; Senn, Walter (2012). Spike-based decision learning of nash equilibria in two-player games. PLoS computational biology, 8(9), pp. 1-12. San Francisco, Calif.: Public Library of Science 10.1371/journal.pcbi.1002691

[img]
Preview
Text
journal.pcbi.1002691.pdf - Published Version
Available under License Creative Commons: Attribution (CC-BY).

Download (587kB) | Preview

Humans and animals face decision tasks in an uncertain multi-agent environment where an agent's strategy may change in time due to the co-adaptation of others strategies. The neuronal substrate and the computational algorithms underlying such adaptive decision making, however, is largely unknown. We propose a population coding model of spiking neurons with a policy gradient procedure that successfully acquires optimal strategies for classical game-theoretical tasks. The suggested population reinforcement learning reproduces data from human behavioral experiments for the blackjack and the inspector game. It performs optimally according to a pure (deterministic) and mixed (stochastic) Nash equilibrium, respectively. In contrast, temporal-difference(TD)-learning, covariance-learning, and basic reinforcement learning fail to perform optimally for the stochastic strategy. Spike-based population reinforcement learning, shown to follow the stochastic reward gradient, is therefore a viable candidate to explain automated decision learning of a Nash equilibrium in two-player games.

Item Type:

Journal Article (Original Article)

Division/Institute:

04 Faculty of Medicine > Pre-clinic Human Medicine > Institute of Physiology
10 Strategic Research Centers > Center for Cognition, Learning and Memory (CCLM)

UniBE Contributor:

Friedrich, Johannes and Senn, Walter

Subjects:

600 Technology > 610 Medicine & health

ISSN:

1553-734X

Publisher:

Public Library of Science

Submitter:

Factscience Import

Date Deposited:

04 Oct 2013 14:35

Last Modified:

20 Jun 2016 09:51

Publisher DOI:

10.1371/journal.pcbi.1002691

PubMed ID:

23028289

Web of Science ID:

000309510900028

BORIS DOI:

10.7892/boris.14214

URI:

https://boris.unibe.ch/id/eprint/14214 (FactScience: 221077)

Actions (login required)

Edit item Edit item
Provide Feedback