Presentation 2014-06-19
Studies on reward mechanisms for self-fulfilling declaration
Masaaki OKA, Taiki TODO, Yuko SAKURAI, Makoto YOKOO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We investigate a mechanism for giving a right incentive to each agent to predict her own future actions and to declare her predictions truthfully. Obtaining an accurate prediction of customer actions is valuable for certain providers who are required to meet customer demands. If an agent predicts an external event that she cannot control herself, any proper scoring rule can give a right incentive. In our problem setting, an agent needs to predict her own action that she can control to maximize her utility. Also, her (gross) utility can vary based on an eternal event. While proper scoring rules have been proposed for eliciting truthful subjective beliefs/prediction of agents in a prediction market, this work presents a new and interesting application for proper scoring rule. We prove that a mechanism that utilizes any strictly proper scoring rule is only a mechanism that can satisfy our goal, assuming an agent can find an optimal declaration that maximizes her expected utility.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Multi-agent system / mechanism design / prediction / scoring rule
Paper # AI2014-8
Date of Issue

Conference Information
Committee AI
Conference Date 2014/6/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Artificial Intelligence and Knowledge-Based Processing (AI)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Studies on reward mechanisms for self-fulfilling declaration
Sub Title (in English)
Keyword(1) Multi-agent system
Keyword(2) mechanism design
Keyword(3) prediction
Keyword(4) scoring rule
1st Author's Name Masaaki OKA
1st Author's Affiliation Kyushu University()
2nd Author's Name Taiki TODO
2nd Author's Affiliation Kyushu University
3rd Author's Name Yuko SAKURAI
3rd Author's Affiliation Kyushu University
4th Author's Name Makoto YOKOO
4th Author's Affiliation Kyushu University
Date 2014-06-19
Paper # AI2014-8
Volume (vol) vol.114
Number (no) 89
Page pp.pp.-
#Pages 6
Date of Issue