Balancing MCTS by Dynamically Adjusting Komi Value

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F11%3A10102972" target="_blank" >RIV/00216208:11320/11:10102972 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Balancing MCTS by Dynamically Adjusting Komi Value

Popis výsledku v původním jazyce

The Monte Carlo Tree Search in the game of Go tends to produce unstable and unreasonable results when used in situations of extreme advantage or disadvantage, due to poor move selection because of low signal-to-noise ratio; notably, this occurs when playing in high handicap games, burdening the computer with further disadvantage against the strong human opponent. We explore and compare multiple approaches to mitigate this problem by artificially evening out the game based on modi- fication of the finalgame score by variable amount of points ("dynamic komi") before storing the result in the game tree. We also compare performance of MCTS and traditional tree search in the context of extreme situations and measure the effect of dynamic komi on actual playing strength of a state-of-art MCTS Go program. Based on our results, we also conjencture on resilience of the game search tree to changes in the evaluation function throughout the search.

Název v anglickém jazyce

Balancing MCTS by Dynamically Adjusting Komi Value

Popis výsledku anglicky

The Monte Carlo Tree Search in the game of Go tends to produce unstable and unreasonable results when used in situations of extreme advantage or disadvantage, due to poor move selection because of low signal-to-noise ratio; notably, this occurs when playing in high handicap games, burdening the computer with further disadvantage against the strong human opponent. We explore and compare multiple approaches to mitigate this problem by artificially evening out the game based on modi- fication of the finalgame score by variable amount of points ("dynamic komi") before storing the result in the game tree. We also compare performance of MCTS and traditional tree search in the context of extreme situations and measure the effect of dynamic komi on actual playing strength of a state-of-art MCTS Go program. Based on our results, we also conjencture on resilience of the game search tree to changes in the evaluation function throughout the search.

Druh

J_x - Nezařazeno - Článek v odborném periodiku (Jimp, Jsc a Jost)

CEP obor

IN - Informatika

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2011

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

ICGA Journal

ISSN

1389-6911

e-ISSN

—

Svazek periodika

34

Číslo periodika v rámci svazku

3

Stát vydavatele periodika

NL - Nizozemsko

Počet stran výsledku

9

Strana od-do

131-139

Kód UT WoS článku

000297955100001

EID výsledku v databázi Scopus

—