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

Imitation learning in a high-dimensional environment is challenging. Mostinverse reinforcement learning (IRL) methods fail to outperform thedemonstrator in such a high-dimensional environment, e.g., Atari domain. Toaddress this challenge, we propose a novel reward learning module to generateintrinsic reward signals via a generative model. Our generative method canperform better forward state transition and backward action encoding, whichimproves the module's dynamics modeling ability in the environment. Thus, ourmodule provides the imitation agent both the intrinsic intention of thedemonstrator and a better exploration ability, which is critical for the agentto outperform the demonstrator. Empirical results show that our methodoutperforms state-of-the-art IRL methods on multiple Atari games, even withone-life demonstration. Remarkably, our method achieves performance that is upto 5 times the performance of the demonstration.

Intrinsic Reward Driven Imitation Learning via Generative Model