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

Offline reinforcement learning (RL) can learn control policies from staticdatasets but, like standard RL methods, it requires reward annotations forevery transition. In many cases, labeling large datasets with rewards may becostly, especially if those rewards must be provided by human labelers, whilecollecting diverse unlabeled data might be comparatively inexpensive. How canwe best leverage such unlabeled data in offline RL? One natural solution is tolearn a reward function from the labeled data and use it to label the unlabeleddata. In this paper, we find that, perhaps surprisingly, a much simpler methodthat simply applies zero rewards to unlabeled data leads to effective datasharing both in theory and in practice, without learning any reward model atall. While this approach might seem strange (and incorrect) at first, weprovide extensive theoretical and empirical analysis that illustrates how ittrades off reward bias, sample complexity and distributional shift, oftenleading to good results. We characterize conditions under which this simplestrategy is effective, and further show that extending it with a simplereweighting approach can further alleviate the bias introduced by usingincorrect reward labels. Our empirical evaluation confirms these findings insimulated robotic locomotion, navigation, and manipulation settings.

How to Leverage Unlabeled Data in Offline Reinforcement Learning