Temporal Difference Learning of Position Evaluation in the Game of Go

Schraudolph, NN; Dayan, P; Sejnowski, TJ

Local TagsRelease HistoryDetailsSummary

Temporal Difference Learning of Position Evaluation in the Game of Go

Schraudolph, N., Dayan, P., & Sejnowski, T. (1994). Temporal Difference Learning of Position Evaluation in the Game of Go. In J. Cowan, G. Tesauro, & J. Alspector (Eds.), Advances in Neural Information Processing Systems 6 (pp. 817-824). San Mateo, CA, USA: Morgan Kaufmann.

Item is Released

show all hide all

Basic

show hide

Item Permalink: https://hdl.handle.net/21.11116/0000-0002-E233-5 Version Permalink: https://hdl.handle.net/21.11116/0000-0005-1A33-5

Genre: Conference Paper

Files

show Files

Locators

show

hide

Locator:
https://papers.nips.cc/paper/820-temporal-difference-learning-of-position-evaluation-in-the-game-of-go.pdf (Publisher version) Open Access status unknown

Description:
-

OA-Status:

Creators

show

hide

Creators:
Schraudolph, NN, Author
Dayan, P¹, Author
Sejnowski, TJ, Author

Affiliations:
1External Organizations, ou_persistent22

Content

show

hide

Free keywords: -

Abstract: The game of Go has a high branching factor that defeats the tree search approach used in computer chess, and long-range spatiotemporal interactions that make position evaluation extremely difficult. Development of conventional Go programs is hampered by their knowledge-intensive nature. We demonstrate a viable alternative by training networks to evaluate Go positions via temporal difference (TD) learning. Our approach is based on network architectures that reflect the spatial organization of both input and reinforcement signals on the Go board, and training protocols that provide exposure to competent (though unlabelled) play. These techniques yield far better performance than undifferentiated networks trained by selfplay alone. A network with less than 500 weights learned within 3,000 games of 9x9 Go a position evaluation function that enables a primitive one-ply search to defeat a commercial Go program at a low playing level.

Details

show

hide

Language(s):

Dates: Date issued: 1994

Publication Status: Issued

Pages: -

Publishing info: -

Table of Contents: -

Rev. Type: -

Identifiers: -

Degree: -

Event

show

hide

Title: Seventh Annual Conference on Neural Information Processing Systems (NIPS 1993)

Place of Event: Denver, CO, USA

Start-/End Date: 1993-11-29 - 1993-12-02

Legal Case

show

Project information

show

Source 1

show

hide

Title: Advances in Neural Information Processing Systems 6

Source Genre: Proceedings

Creator(s):
Cowan, JD, Editor
Tesauro, G, Editor
Alspector, J, Editor

Affiliations:
-

Publ. Info: San Mateo, CA, USA : Morgan Kaufmann

Pages: - Volume / Issue: - Sequence Number: - Start / End Page: 817 - 824 Identifier: ISBN: 1-55860-322-0