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

We consider the problem of how a teacher algorithm can enable an unknown DeepReinforcement Learning (DRL) student to become good at a skill over a widerange of diverse environments. To do so, we study how a teacher algorithm canlearn to generate a learning curriculum, whereby it sequentially samplesparameters controlling a stochastic procedural generation of environments.Because it does not initially know the capacities of its student, a keychallenge for the teacher is to discover which environments are easy, difficultor unlearnable, and in what order to propose them to maximize the efficiency oflearning over the learnable ones. To achieve this, this problem is transformedinto a surrogate continuous bandit problem where the teacher samplesenvironments in order to maximize absolute learning progress of its student. Wepresent a new algorithm modeling absolute learning progress with Gaussianmixture models (ALP-GMM). We also adapt existing algorithms and provide acomplete study in the context of DRL. Using parameterized variants of theBipedalWalker environment, we study their efficiency to personalize a learningcurriculum for different learners (embodiments), their robustness to the ratioof learnable/unlearnable environments, and their scalability to non-linear andhigh-dimensional parameter spaces. Videos and code are available athttps://github.com/flowersteam/teachDeepRL.

Teacher algorithms for curriculum learning of Deep RL in continuously parameterized environments