When in Doubt, Think Slow: Iterative Reasoning with Latent Imagination

arXiv:2402.15283v1 Announce Type: new
Abstract: In an unfamiliar setting, a model-based reinforcement learning agent can be limited by the accuracy of its world model. In this work, we present a novel, training-free approach to improving the performance of such agents separately from planning and learning. We do so by applying iterative inference at decision-time, to fine-tune the inferred agent states based on the coherence of future state representations. Our approach achieves a consistent improvement in both reconstruction accuracy and task …

abstract accuracy agent agents arxiv cs.ai cs.lg free imagination iterative novel performance planning reasoning reinforcement reinforcement learning think training type work world