Compact Representation of Value Function in Partially Observable Stochastic Games

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F19%3A00332701" target="_blank" >RIV/68407700:21230/19:00332701 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.ijcai.org/proceedings/2019/50" target="_blank" >https://www.ijcai.org/proceedings/2019/50</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.24963/ijcai.2019/50" target="_blank" >10.24963/ijcai.2019/50</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Compact Representation of Value Function in Partially Observable Stochastic Games

Popis výsledku v původním jazyce

Value methods for solving stochastic games with partial observability model the uncertainty of the players as a probability distribution over possible states, where the dimension of the belief space is the number of states. For many practical problems, there are exponentially many states which causes scalability problems. We propose an abstraction technique that addresses this curse of dimensionality by projecting the high-dimensional beliefs onto characteristic vectors of significantly lower dimension (e.g., marginal probabilities). Our main contributions are (1) a novel compact representation of the uncertainty in partially observable stochastic games and (2) a novel algorithm using this representation that is based on existing state-of-the-art algorithms for solving stochastic games with partial observability. Experimental evaluation confirms that the new algorithm using the compact representation dramatically increases scalability compared to the state of the art.

Název v anglickém jazyce

Compact Representation of Value Function in Partially Observable Stochastic Games

Popis výsledku anglicky

Value methods for solving stochastic games with partial observability model the uncertainty of the players as a probability distribution over possible states, where the dimension of the belief space is the number of states. For many practical problems, there are exponentially many states which causes scalability problems. We propose an abstraction technique that addresses this curse of dimensionality by projecting the high-dimensional beliefs onto characteristic vectors of significantly lower dimension (e.g., marginal probabilities). Our main contributions are (1) a novel compact representation of the uncertainty in partially observable stochastic games and (2) a novel algorithm using this representation that is based on existing state-of-the-art algorithms for solving stochastic games with partial observability. Experimental evaluation confirms that the new algorithm using the compact representation dramatically increases scalability compared to the state of the art.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GJ19-24384Y" target="_blank" >GJ19-24384Y: Výpočet rovnovážných strategií v dynamických hrách</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2019

Kód důvěrnosti údajů

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

Název statě ve sborníku

Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence

ISBN

978-0-9992411-4-1

ISSN

—

e-ISSN

1045-0823

Počet stran výsledku

7

Strana od-do

350-356

Název nakladatele

International Joint Conferences on Artificial Intelligence Organization

Místo vydání

—

Místo konání akce

Macau

Datum konání akce

10. 8. 2019

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—