A unifying learning framework for building artificial game-playing agents

NCUIR > Office of Teaching Centers > Language Center > journal & Dissertation > Item 987654321/107856

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/107856

Title:	A unifying learning framework for building artificial game-playing agents
Authors:	陳雯玲;Chen, Wenlin;Chen, Yixin;Levine, David K.
Contributors:	總教學中心語言中心
Keywords:	Agents (artificial intelligence);Algorithms;Artificial Intelligence;Bayesian analysis;Buildings;Complex Systems;Computer Science;Construction;Construction costs;Economics;Games;Learning;Machine learning;Mathematical models;Mathematics
Date:	2015-01-31
Issue Date:	2026-04-23 14:27:27 (UTC+8)
Publisher:	Springer Netherlands;Cham: Springer International Publishing
Abstract:	摘要： This paper investigates learning-based agents that are capable of mimicking human behavior in game playing, a central task in computational economics. Although computational economists have developed various game-playing agents, well-established machine learning methods such as graphical models have not been applied before. Leveraging probabilistic graphical models, this paper presents a novel sequential Bayesian network (SBN) framework for building artificial game-playing agents. We show that many existing agents, including reinforcement learning, fictitious play, and many of their variants, have a unified Bayesian explanation within the proposed SBN framework. Moreover, we discover that SBN can handle various important settings of game playing, allowing for a broad scope of its use in economics. SBN not only provides a unifying and satisfying framework to explain existing learning approaches in virtual economies, but also enables the development of new algorithms that are stronger or have fewer restrictions. In this paper, we derive a new algorithm, Hidden Markovian Play (HMP), from the generic SBN model to handle an important but difficult setting in which a player cannot observe the opponent’s strategy and payoff. It leverages Markovian learning to infer unobservable information, leading to higher quality of the agents. Experiments on real-world field experiments in evaluating economies show that our HMP model outperforms the baseline algorithms for building artificial agents. 其他題名： Ann Math Artif Intell 出版者： Cham: Springer International Publishing 出版日期： 2015-04-01 出處： Annals of mathematics and artificial intelligence, 2015-04, Vol.73 (3-4), p.335-358 版權： Springer International Publishing Switzerland 2015 版權： Springer International Publishing Switzerland 2015. 識別號： ISSN: 1012-2443 識別號： EISSN: 1573-7470 識別號： DOI: 10.1007/s10472-015-9450-1
Appears in Collections:	[Language Center ] journal & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	41	View/Open

社群 sharing

Loading...