Learning Social Preferences in Games

Yakov Gal, Avi Pfeffer, Francesca Marzo, Barbara J. Grosz

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents a machine-learning approach to modeling human behavior in one-shot games. It provides a framework for representing and reasoning about the social factors that affect people’s play. The model predicts how a human player is likely to react to different actions of another player, and these predictions are used to determine the best possible strategy for that player. Data collection and evaluation of the model were performed on a negotiation game in which humans played against each other and against computer models playing various strategies. A computer player trained on human data outplayed Nash equilibrium and Nash bargaining computer players as well as humans. It also generalized to play people and game situations it had not seen before.
Original languageEnglish
Title of host publicationPROCEEDINGS OF THE NINETEENTH NATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE
Place of PublicationSan Jose, California, USA
PublisherAAAI Press
Pages226-231
Number of pages6
Publication statusPublished - 2004
EventNINETEENTH NATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE - San Jose, United States
Duration: 25 Jul 200429 Jul 2004

Conference

ConferenceNINETEENTH NATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE
Abbreviated titleAAAI-04
Country/TerritoryUnited States
CitySan Jose
Period25/07/0429/07/04

Fingerprint

Dive into the research topics of 'Learning Social Preferences in Games'. Together they form a unique fingerprint.

Cite this