Training an Assassin AI for The Resistance: Avalon
This work addresses the underdeveloped area of AI for social deduction games, though it is incremental as it focuses on a single phase and role.
The authors tackled the problem of creating an AI for the Assassin role in The Resistance: Avalon, a partially observable social deduction game, by training classifiers on a public dataset and achieved above-average human performance using a linear support vector classifier.
The Resistance: Avalon is a partially observable social deduction game. This area of AI game playing is fairly undeveloped. Implementing an AI for this game involves multiple components specific to each phase as well as role in the game. In this paper, we plan to iteratively develop the required components for each role/phase by first addressing the Assassination phase which can be modeled as a machine learning problem. Using a publicly available dataset from an online version of the game, we train classifiers that emulate an Assassin. After trying various classification techniques, we are able to achieve above average human performance using a simple linear support vector classifier. The eventual goal of this project is to pursue developing an intelligent and complete Avalon player that can play through each phase of the game as any role.