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

Interactive reinforcement learning has shown promise in learning complexrobotic tasks. However, the process can be human-intensive due to therequirement of a large amount of interactive feedback. This paper presents anew method that uses scores provided by humans instead of pairwise preferencesto improve the feedback efficiency of interactive reinforcement learning. Ourkey insight is that scores can yield significantly more data than pairwisepreferences. Specifically, we require a teacher to interactively score the fulltrajectories of an agent to train a behavioral policy in a sparse rewardenvironment. To avoid unstable scores given by humans negatively impacting thetraining process, we propose an adaptive learning scheme. This enables thelearning paradigm to be insensitive to imperfect or unreliable scores. Weextensively evaluate our method for robotic locomotion and manipulation tasks.The results show that the proposed method can efficiently learn near-optimalpolicies by adaptive learning from scores while requiring less feedbackcompared to pairwise preference learning methods. The source codes are publiclyavailable at https://github.com/SSKKai/Interactive-Scoring-IRL.

Boosting Feedback Efficiency of Interactive Reinforcement Learning by Adaptive Learning from Scores