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

Unsupervised visual representation learning offers the opportunity toleverage large corpora of unlabeled trajectories to form useful visualrepresentations, which can benefit the training of reinforcement learning (RL)algorithms. However, evaluating the fitness of such representations requirestraining RL algorithms which is computationally intensive and has high varianceoutcomes. To alleviate this issue, we design an evaluation protocol forunsupervised RL representations with lower variance and up to 600x lowercomputational cost. Inspired by the vision community, we propose two linearprobing tasks: predicting the reward observed in a given state, and predictingthe action of an expert in a given state. These two tasks are generallyapplicable to many RL domains, and we show through rigorous experimentationthat they correlate strongly with the actual downstream control performance onthe Atari100k Benchmark. This provides a better method for exploring the spaceof pretraining algorithms without the need of running RL evaluations for everysetting. Leveraging this framework, we further improve existing self-supervisedlearning (SSL) recipes for RL, highlighting the importance of the forwardmodel, the size of the visual backbone, and the precise formulation of theunsupervised objective.

Light-weight probing of unsupervised representations for Reinforcement Learning