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

The potential benefits of model-free reinforcement learning to real roboticssystems are limited by its uninformed exploration that leads to slowconvergence, lack of data-efficiency, and unnecessary interactions with theenvironment. To address these drawbacks we propose a method that combinesreinforcement and imitation learning by shaping the reward function with astate-and-action-dependent potential that is trained from demonstration data,using a generative model. We show that this accelerates policy learning byspecifying high-value areas of the state and action space that are worthexploring first. Unlike the majority of existing methods that assume optimaldemonstrations and incorporate the demonstration data as hard constraints onpolicy optimization, we instead incorporate demonstration data as advice in theform of a reward shaping potential trained as a generative model of states andactions. In particular, we examine both normalizing flows and GenerativeAdversarial Networks to represent these potentials. We show that, unlike manyexisting approaches that incorporate demonstrations as hard constraints, ourapproach is unbiased even in the case of suboptimal and noisy demonstrations.We present an extensive range of simulations, as well as experiments on theFranka Emika 7DOF arm, to demonstrate the practicality of our method.

Shaping Rewards for Reinforcement Learning with Imperfect Demonstrations using Generative Models