GAIG Game AI Research Group @ QMUL

Rinascimento: using event-value functions for playing Splendor


Abstract

In the realm of games research, Artificial General Intelligence algorithms often use score as main reward signal for learning or playing actions. However this has shown its severe limitations when the point rewards are very rare or absent until the end of the game. This paper proposes a new approach based on event logging: the game state triggers an event every time one of its features changes. These events are processed by an Event-value Function (EF) that assigns a value to a single action or a sequence. The experiments have shown that such approach can mitigate the problem of scarce point rewards and improve the AI performance. Furthermore this represents a step forward in controlling the strategy adopted by the artificial agent, by describing a much richer and controllable behavioural space through the EF. Tuned EF are able to neatly synthesise the relevance of the events in the game. Agents using an EF show more robust when playing games with several opponents.
URL: https://ieeexplore.ieee.org/document/9231691

Cite this work

@inproceedings{bravi2020rinascimento,
author= {Bravi, Ivan and Lucas, Simon M},
title= {{Rinascimento: using event-value functions for playing Splendor}},
year= {2020},
booktitle= {{IEEE Conference on Games (CoG)}},
pages= {283--290},
url= {https://ieeexplore.ieee.org/document/9231691},
abstract= {In the realm of games research, Artificial General Intelligence algorithms often use score as main reward signal for learning or playing actions. However this has shown its severe limitations when the point rewards are very rare or absent until the end of the game. This paper proposes a new approach based on event logging: the game state triggers an event every time one of its features changes. These events are processed by an Event-value Function (EF) that assigns a value to a single action or a sequence. The experiments have shown that such approach can mitigate the problem of scarce point rewards and improve the AI performance. Furthermore this represents a step forward in controlling the strategy adopted by the artificial agent, by describing a much richer and controllable behavioural space through the EF. Tuned EF are able to neatly synthesise the relevance of the events in the game. Agents using an EF show more robust when playing games with several opponents.},
}

Similar Posts

Comments