Neural Information Processing Systems http://nips.cc/

As a challenging multi-player card game, DouDizhu has recently drawn much attention for analyzing competition and collaboration in imperfect-information games. In this paper, we propose PerfectDou, a state-of-the-art DouDizhu AI system that dominates the game, in an actor-critic framework with a proposed technique named perfect information distillation.

If the reward function does not cover all important aspects of the task, the agent could easily learn undesirable behaviors [4].

First, we rewrite Eq. (3) as max

Learning in real-world multiagent tasks is challenging due to the usual partial observability ofeach agent. Previous efforts alleviate thepartial observability by historical hidden states with Recurrent Neural Networks, however, they do not consider themultiagent characters thateither themultiagent observationconsists ofanumber ofobject entities orthe action space shows clear entity interactions.

Then,wecalculatethe mean stiffness of the value network across all state pairs and report its average computed over all trainingepochs. Eachagentis trained on 200 training levels for 25M environment steps. The mean and standard deviation are computedover10differentruns. Morespecifically,wecollect100 training episodes throughout the training and evaluate the value network prediction for the initial stateofeachtrajectory. Each agent is trained on 200 training levels for 25M environment steps.

Fortraining,PPOfirst collects trajectoriesτ using the old policy networkπθold right before the update.

Neural Information Processing Systems http://nips.cc/