Presentation	1997/5/23 Reinforcement learning algorithm for prediction and control of ball games Shingo Ochiai, Masaki Sano, Yasuji Sawada,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	An interesting question about learning is how an embedded agent call improve performance while acting in complex dynamical environment. Supervised learning is not feasible because precise knowledge of dynamical environment and correct response of agent are not available a priori. As a case study of control problem we choose two types of ball playing games ; shooting in basketball and squash tennis. In both cases, robust and efficient algorithm is needed. We applied the stochastic reinforcement learning method with temporal difference (TD) algorithm for controlling and prediction. The result was successful.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	stochastic reinforcement learning / TD learning / RBF
Paper #	NC97-10
Date of Issue

Conference Information
Committee	NC
Conference Date	1997/5/23(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	Neurocomputing (NC)
Language	JPN
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	Reinforcement learning algorithm for prediction and control of ball games
Sub Title (in English)
Keyword(1)	stochastic reinforcement learning
Keyword(2)	TD learning
Keyword(3)	RBF
1st Author's Name	Shingo Ochiai
1st Author's Affiliation	Graduate School of Information Sciences, TOHOKU University()
2nd Author's Name	Masaki Sano
2nd Author's Affiliation	Graduate School of Information Sciences, TOHOKU University
3rd Author's Name	Yasuji Sawada
3rd Author's Affiliation	Graduate School of Information Sciences, TOHOKU University
Date	1997/5/23
Paper #	NC97-10
Volume (vol)	vol.97
Number (no)	69
Page	pp.pp.-
#Pages	7
Date of Issue