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

Sample efficiency and risk-awareness are central to the development ofpractical reinforcement learning (RL) for complex decision-making. The formercan be addressed by transfer learning and the latter by optimizing some utilityfunction of the return. However, the problem of transferring skills in arisk-aware manner is not well-understood. In this paper, we address the problemof risk-aware policy transfer between tasks in a common domain that differ onlyin their reward functions, in which risk is measured by the variance of rewardstreams. Our approach begins by extending the idea of generalized policyimprovement to maximize entropic utilities, thus extending policy improvementvia dynamic programming to sets of policies and levels of risk-aversion. Next,we extend the idea of successor features (SF), a value function representationthat decouples the environment dynamics from the rewards, to capture thevariance of returns. Our resulting risk-aware successor features (RaSF)integrate seamlessly within the RL framework, inherit the superior taskgeneralization ability of SFs, and incorporate risk-awareness into thedecision-making. Experiments on a discrete navigation domain and control of asimulated robotic arm demonstrate the ability of RaSFs to outperformalternative methods including SFs, when taking the risk of the learned policiesinto account.

Risk-Aware Transfer in Reinforcement Learning using Successor Features