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

Reinforcement Learning (RL) algorithms aim to learn an optimal policy byiteratively sampling actions to learn how to maximize the total expectedreturn, $R(x)$. GFlowNets are a special class of algorithms designed togenerate diverse candidates, $x$, from a discrete set, by learning a policythat approximates the proportional sampling of $R(x)$. GFlowNets exhibitimproved mode discovery compared to conventional RL algorithms, which is veryuseful for applications such as drug discovery and combinatorial search.However, since GFlowNets are a relatively recent class of algorithms, manytechniques which are useful in RL have not yet been associated with them. Inthis paper, we study the utilization of a replay buffer for GFlowNets. Weexplore empirically various replay buffer sampling techniques and assess theimpact on the speed of mode discovery and the quality of the modes discovered.Our experimental results in the Hypergrid toy domain and a molecule synthesisenvironment demonstrate significant improvements in mode discovery whentraining with a replay buffer, compared to training only with trajectoriesgenerated on-policy.

An Empirical Study of the Effectiveness of Using a Replay Buffer on Mode Discovery in GFlowNets