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

Reinforcement learning provides an automated framework for learning behaviorsfrom high-level reward specifications, but in practice the choice of rewardfunction can be crucial for good results -- while in principle the reward onlyneeds to specify what the task is, in reality practitioners often need todesign more detailed rewards that provide the agent with some hints about howthe task should be completed. The idea of this type of ``reward-shaping'' hasbeen often discussed in the literature, and is often a critical part ofpractical applications, but there is relatively little formal characterizationof how the choice of reward shaping can yield benefits in sample complexity. Inthis work, we build on the framework of novelty-based exploration to provide asimple scheme for incorporating shaped rewards into RL along with an analysistool to show that particular choices of reward shaping provably improve sampleefficiency. We characterize the class of problems where these gains areexpected to be significant and show how this can be connected to practicalalgorithms in the literature. We confirm that these results hold in practice inan experimental evaluation, providing an insight into the mechanisms throughwhich reward shaping can significantly improve the complexity of reinforcementlearning while retaining asymptotic performance.

Unpacking Reward Shaping: Understanding the Benefits of Reward Engineering on Sample Complexity