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

Offline goal-conditioned RL (GCRL) offers a feasible paradigm to learngeneral-purpose policies from diverse and multi-task offline datasets. Despitenotable recent progress, the predominant offline GCRL methods have beenrestricted to model-free approaches, constraining their capacity to tacklelimited data budgets and unseen goal generalization. In this work, we propose anovel two-stage model-based framework, Goal-conditioned Offline Planning(GOPlan), including (1) pretraining a prior policy capable of capturingmulti-modal action distribution within the multi-goal dataset; (2) employingthe reanalysis method with planning to generate imagined trajectories forfunetuning policies. Specifically, the prior policy is based on anadvantage-weighted Conditioned Generative Adversarial Networks that exhibitsdistinct mode separation to overcome the pitfalls of out-of-distribution (OOD)actions. For further policy optimization, the reanalysis method generateshigh-quality imaginary data by planning with learned models for bothintra-trajectory and inter-trajectory goals. Through experimental evaluations,we demonstrate that GOPlan achieves state-of-the-art performance on variousoffline multi-goal manipulation tasks. Moreover, our results highlight thesuperior ability of GOPlan to handle small data budgets and generalize to OODgoals.

GOPlan: Goal-conditioned Offline Reinforcement Learning by Planning with Learned Models