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

Bayesian experimental design (BED) is to answer the question that how tochoose designs that maximize the information gathering. For implicit models,where the likelihood is intractable but sampling is possible, conventional BEDmethods have difficulties in efficiently estimating the posterior distributionand maximizing the mutual information (MI) between data and parameters. Recentwork proposed the use of gradient ascent to maximize a lower bound on MI todeal with these issues. However, the approach requires a sampling path tocompute the pathwise gradient of the MI lower bound with respect to the designvariables, and such a pathwise gradient is usually inaccessible for implicitmodels. In this paper, we propose a novel approach that leverages recentadvances in stochastic approximate gradient ascent incorporated with a smoothedvariational MI estimator for efficient and robust BED. Without the necessity ofpathwise gradients, our approach allows the design process to be achievedthrough a unified procedure with an approximate gradient for implicit models.Several experiments show that our approach outperforms baseline methods, andsignificantly improves the scalability of BED in high-dimensional problems.

A Scalable Gradient-Free Method for Bayesian Experimental Design with Implicit Models