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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Diffusion-based planning has shown promising results in long-horizon,sparse-reward tasks by training trajectory diffusion models and conditioningthe sampled trajectories using auxiliary guidance functions. However, due totheir nature as generative models, diffusion models are not guaranteed togenerate feasible plans, resulting in failed execution and precluding plannersfrom being useful in safety-critical applications. In this work, we propose anovel approach to refine unreliable plans generated by diffusion models byproviding refining guidance to error-prone plans. To this end, we suggest a newmetric named restoration gap for evaluating the quality of individual plansgenerated by the diffusion model. A restoration gap is estimated by a gappredictor which produces restoration gap guidance to refine a diffusionplanner. We additionally present an attribution map regularizer to preventadversarial refining guidance that could be generated from the sub-optimal gappredictor, which enables further refinement of infeasible plans. We demonstratethe effectiveness of our approach on three different benchmarks in offlinecontrol settings that require long-horizon planning. We also illustrate thatour approach presents explainability by presenting the attribution maps of thegap predictor and highlighting error-prone transitions, allowing for a deeperunderstanding of the generated plans.

Refining Diffusion Planner for Reliable Behavior Synthesis by Automatic Detection of Infeasible Plans