Policy-shaped prediction: avoiding distractions in model-based reinforcement learning
This addresses a known vulnerability in MBRL for scenarios with irrelevant predictable content, offering an incremental improvement towards more robust learning.
The paper tackles the problem of model-based reinforcement learning (MBRL) being distracted by predictable but irrelevant world details, which can degrade policy learning. They introduced a method combining pretrained segmentation, task-aware reconstruction loss, and adversarial learning, outperforming other approaches in a novel environment with intricate background distractions.
Model-based reinforcement learning (MBRL) is a promising route to sample-efficient policy optimization. However, a known vulnerability of reconstruction-based MBRL consists of scenarios in which detailed aspects of the world are highly predictable, but irrelevant to learning a good policy. Such scenarios can lead the model to exhaust its capacity on meaningless content, at the cost of neglecting important environment dynamics. While existing approaches attempt to solve this problem, we highlight its continuing impact on leading MBRL methods -- including DreamerV3 and DreamerPro -- with a novel environment where background distractions are intricate, predictable, and useless for planning future actions. To address this challenge we develop a method for focusing the capacity of the world model through synergy of a pretrained segmentation model, a task-aware reconstruction loss, and adversarial learning. Our method outperforms a variety of other approaches designed to reduce the impact of distractors, and is an advance towards robust model-based reinforcement learning.