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

Designing a distribution of environments in which RL agents can learninteresting and useful skills is a challenging and poorly understood task, formulti-agent environments the difficulties are only exacerbated. One approach isto train a second RL agent, called a teacher, who samples environments that areconducive for the learning of student agents. However, most previous proposalsfor teacher rewards do not generalize straightforwardly to the multi-agentsetting. We examine a set of intrinsic teacher rewards derived from predictionproblems that can be applied in multi-agent settings and evaluate them inMujoco tasks such as multi-agent Hide and Seek as well as a diagnosticsingle-agent maze task. Of the intrinsic rewards considered we found valuedisagreement to be most consistent across tasks, leading to faster and morereliable emergence of advanced skills in Hide and Seek and the maze task.Another candidate intrinsic reward considered, value prediction error, alsoworked well in Hide and Seek but was susceptible to noisy-TV style distractionsin stochastic environments. Policy disagreement performed well in the maze taskbut did not speed up learning in Hide and Seek. Our results suggest thatintrinsic teacher rewards, and in particular value disagreement, are apromising approach for automating both single and multi-agent environmentdesign.

AutoDIME: Automatic Design of Interesting Multi-Agent Environments