World-Model Policy Arbiter for Goal-Conditioned Reinforcement Learning
The World-Model Policy Arbiter (WMPA) is a test-time framework for offline goal-conditioned reinforcement learning that selects the best policy from a bank of frozen policies. It works by rolling each policy forward in a learned world model, scoring imagined futures with a shared value function, and executing the highest-scoring one for short intervals. This improves macro success rates significantly.