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

As more non-AI experts use complex AI systems for daily tasks, there has beenan increasing effort to develop methods that produce explanations of AIdecision making that are understandable by non-AI experts. Towards this effort,leveraging higher-level concepts and producing concept-based explanations havebecome a popular method. Most concept-based explanations have been developedfor classification techniques, and we posit that the few existing methods forsequential decision making are limited in scope. In this work, we firstcontribute a desiderata for defining concepts in sequential decision makingsettings. Additionally, inspired by the Protege Effect which states explainingknowledge often reinforces one's self-learning, we explore how concept-basedexplanations of an RL agent's decision making can in turn improve the agent'slearning rate, as well as improve end-user understanding of the agent'sdecision making. To this end, we contribute a unified framework,State2Explanation (S2E), that involves learning a joint embedding model betweenstate-action pairs and concept-based explanations, and leveraging such learnedmodel to both (1) inform reward shaping during an agent's training, and (2)provide explanations to end-users at deployment for improved task performance.Our experimental validations, in Connect 4 and Lunar Lander, demonstrate thesuccess of S2E in providing a dual-benefit, successfully informing rewardshaping and improving agent learning rate, as well as significantly improvingend user task performance at deployment time.

State2Explanation: Concept-Based Explanations to Benefit Agent Learning and User Understanding