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

Recently, sequence learning methods have been applied to the problem ofoff-policy Reinforcement Learning, including the seminal work on DecisionTransformers, which employs transformers for this task. Since transformers areparameter-heavy, cannot benefit from history longer than a fixed window size,and are not computed using recurrence, we set out to investigate thesuitability of the S4 family of models, which are based on state-space layersand have been shown to outperform transformers, especially in modelinglong-range dependencies. In this work we present two main algorithms: (i) anoff-policy training procedure that works with trajectories, while stillmaintaining the training efficiency of the S4 model. (ii) An on-policy trainingprocedure that is trained in a recurrent manner, benefits from long-rangedependencies, and is based on a novel stable actor-critic mechanism. Ourresults indicate that our method outperforms multiple variants of decisiontransformers, as well as the other baseline methods on most tasks, whilereducing the latency, number of parameters, and training time by several ordersof magnitude, making our approach more suitable for real-world RL.

Decision S4: Efficient Sequence-Based RL via State Spaces Layers