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

A key challenge for reinforcement learning is solving long-horizon planningproblems. Recent work has leveraged programs to guide reinforcement learning inthese settings. However, these approaches impose a high manual burden on theuser since they must provide a guiding program for every new task. Partiallyobserved environments further complicate the programming task because theprogram must implement a strategy that correctly, and ideally optimally,handles every possible configuration of the hidden regions of the environment.We propose a new approach, model predictive program synthesis (MPPS), that usesprogram synthesis to automatically generate the guiding programs. It trains agenerative model to predict the unobserved portions of the world, and thensynthesizes a program based on samples from this model in a way that is robustto its uncertainty. In our experiments, we show that our approach significantlyoutperforms non-program-guided approaches on a set of challenging benchmarks,including a 2D Minecraft-inspired environment where the agent must complete acomplex sequence of subtasks to achieve its goal, and achieves a similarperformance as using handcrafted programs to guide the agent. Our resultsdemonstrate that our approach can obtain the benefits of program-guidedreinforcement learning without requiring the user to provide a new guidingprogram for every new task.

Program Synthesis Guided Reinforcement Learning for Partially Observed Environments