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

Explaining the behavior of reinforcement learning agents operating insequential decision-making settings is challenging, as their behavior isaffected by a dynamic environment and delayed rewards. Methods that help usersunderstand the behavior of such agents can roughly be divided into localexplanations that analyze specific decisions of the agents and globalexplanations that convey the general strategy of the agents. In this work, westudy a novel combination of local and global explanations for reinforcementlearning agents. Specifically, we combine reward decomposition, a localexplanation method that exposes which components of the reward functioninfluenced a specific decision, and HIGHLIGHTS, a global explanation methodthat shows a summary of the agent's behavior in decisive states. We conductedtwo user studies to evaluate the integration of these explanation methods andtheir respective benefits. Our results show significant benefits for bothmethods. In general, we found that the local reward decomposition was moreuseful for identifying the agents' priorities. However, when there was only aminor difference between the agents' preferences, then the global informationprovided by HIGHLIGHTS additionally improved participants' understanding.

Integrating Policy Summaries with Reward Decomposition for Explaining Reinforcement Learning Agents