Learning to Generalize with Object-centric Agents in the Open World Survival Game Crafter. (arXiv:2208.03374v1 [cs.LG])

Reinforcement learning agents must generalize beyond their training
experience. Prior work has focused mostly on identical training and evaluation
environments. Starting from the recently introduced Crafter benchmark, a 2D
open world survival game, we introduce a new set of environments suitable for
evaluating some agent's ability to generalize on previously unseen (numbers of)
objects and to adapt quickly (meta-learning). In Crafter, the agents are
evaluated by the number of unlocked achievements (such as collecting resources)
when trained for 1M steps. …

agents arxiv game learning lg survival