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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Neural bandits have been shown to provide an efficient solution to practicalsequential decision tasks that have nonlinear reward functions. The maincontributor to that success is approximate Bayesian inference, which enablesneural network (NN) training with uncertainty estimates. However, Bayesian NNsoften suffer from a prohibitive computational overhead or operate on a subsetof parameters. Alternatively, certain classes of infinite neural networks wereshown to directly correspond to Gaussian processes (GP) with neural kernels(NK). NK-GPs provide accurate uncertainty estimates and can be trained fasterthan most Bayesian NNs. We propose to guide common bandit policies with NKdistributions and show that NK bandits achieve state-of-the-art performance onnonlinear structured data. Moreover, we propose a framework for measuringindependently the ability of a bandit algorithm to learn representations andexplore, and use it to analyze the impact of NK distributions w.r.t.~those twoaspects. We consider policies based on a GP and a Student's t-process (TP).Furthermore, we study practical considerations, such as training frequency andmodel partitioning. We believe our work will help better understand the impactof utilizing NKs in applied settings.

Empirical analysis of representation learning and exploration in neural kernel bandits