AI ChronicleAI 纪年 Full timeline →

Research breakthrough · curated entry

TD-Gammon reaches backgammon mastery by self-play

TD-Gammon

Date
June 1992
Category
Research breakthrough
Organization / Model
IBM
Evidence
2 sources · 2 evidence families

Tesauro let a network learn position evaluation through self-play and temporal-difference learning, approaching human champion strength. It showed reinforcement learning could work in complex decision tasks without hand-built features, foreshadowing the AlphaGo route.

Curator's note: Cross-checked against the original paper and the Wikipedia article.

Curated sources · 2

  1. Original paper · Communications of the ACM1995-03

    TD-Gammon

    doi.org

  2. Reference material · Wikipedia1992-06-01

    TD-Gammon

    en.wikipedia.org

Cite this page

This page is a curated static snapshot and can be cited directly. Preferred format GB/T 7714—2025 (electronic resource [EB/OL]); the other 5 styles open below, together with the machine-readable files.

GB/T 7714—2025 (electronic resource [EB/OL])

AI 纪年. TD-Gammon 用神经网络与强化学习达到西洋双陆棋顶尖水平[EB/OL]. (1992-06)[2026-11-03]. https://aihist.net/e/td-gammon-1992.html.

APA 7 · American Psychological Association

AI 纪年. (1992, June). TD-Gammon 用神经网络与强化学习达到西洋双陆棋顶尖水平 [Web page]. https://aihist.net/e/td-gammon-1992.html

MLA 9 · Modern Language Association

"TD-Gammon 用神经网络与强化学习达到西洋双陆棋顶尖水平." AI 纪年, June 1992, aihist.net/e/td-gammon-1992.html. Accessed 3 Nov. 2026.

Chicago 17 · notes-bibliography

AI 纪年. "TD-Gammon 用神经网络与强化学习达到西洋双陆棋顶尖水平." Published June 1992. Accessed Nov. 3, 2026. https://aihist.net/e/td-gammon-1992.html.

IEEE · citation-reference list

"TD-Gammon 用神经网络与强化学习达到西洋双陆棋顶尖水平," AI 纪年, June 1992. [Online]. Available: https://aihist.net/e/td-gammon-1992.html. [Accessed: 3 Nov. 2026].

ACM · author-date variant

AI 纪年. June 1992. TD-Gammon 用神经网络与强化学习达到西洋双陆棋顶尖水平. Retrieved November 3, 2026 from https://aihist.net/e/td-gammon-1992.html.

All six styles come from one template and can be compared word for word. [引用日期], Accessed, urldate and Y2 all carry this page's build date, 2026-11-03 — a static page is an immutable snapshot, so substitute the date you actually consulted it. The address in every style is the archive's canonical page: this English page translates the same snapshot, and the work keeps its name, so one record cites both. For one-click copying use the same entry in the timeline, which takes the day you opened it. Expand the other styles before printing.