English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Visual model-based RL methods typically encode image observations intolow-dimensional representations in a manner that does not eliminate redundantinformation. This leaves them susceptible to spurious variations -- changes intask-irrelevant components such as background distractors or lightingconditions. In this paper, we propose a visual model-based RL method thatlearns a latent representation resilient to such spurious variations. Ourtraining objective encourages the representation to be maximally predictive ofdynamics and reward, while constraining the information flow from theobservation to the latent representation. We demonstrate that this objectivesignificantly bolsters the resilience of visual model-based RL methods tovisual distractors, allowing them to operate in dynamic environments. We thenshow that while the learned encoder is resilient to spirious variations, it isnot invariant under significant distribution shift. To address this, we proposea simple reward-free alignment procedure that enables test time adaptation ofthe encoder. This allows for quick adaptation to widely differing environmentswithout having to relearn the dynamics and policy. Our effort is a step towardsmaking model-based RL a practical and useful tool for dynamic, diverse domains.We show its effectiveness in simulation benchmarks with significant spuriousvariations as well as a real-world egocentric navigation task with noisy TVs inthe background. Videos and code at https://zchuning.github.io/repo-website/.

RePo: Resilient Model-Based Reinforcement Learning by Regularizing Posterior Predictability