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, deep reinforcement learning (RL) has achieved remarkable empiricalsuccess by integrating deep neural networks into RL frameworks. However, thesealgorithms often require a large number of training samples and admit littletheoretical understanding. To mitigate these issues, we propose a theoreticallyprincipled nearest neighbor (NN) function approximator that can improve thevalue networks in deep RL methods. Inspired by human similarity judgments, theNN approximator estimates the action values using rollouts on past observationsand can provably obtain a small regret bound that depends only on the intrinsiccomplexity of the environment. We present (1) Nearest Neighbor Actor-Critic(NNAC), an online policy gradient algorithm that demonstrates the practicalityof combining function approximation with deep RL, and (2) a plug-and-play NNupdate module that aids the training of existing deep RL methods. Experimentson classical control and MuJoCo locomotion tasks show that the NN-acceleratedagents achieve higher sample efficiency and stability than the baseline agents.Based on its theoretical benefits, we believe that the NN approximator can befurther applied to other complex domains to speed-up learning.

Theoretically Principled Deep RL Acceleration via Nearest Neighbor Function Approximation