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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Standard MCMC methods can scale poorly to big data settings due to the needto evaluate the likelihood at each iteration. There have been a number ofapproximate MCMC algorithms that use sub-sampling ideas to reduce thiscomputational burden, but with the drawback that these algorithms no longertarget the true posterior distribution. We introduce a new family of MonteCarlo methods based upon a multi-dimensional version of the Zig-Zag process of(Bierkens, Roberts, 2017), a continuous time piecewise deterministic Markovprocess. While traditional MCMC methods are reversible by construction (aproperty which is known to inhibit rapid convergence) the Zig-Zag processoffers a flexible non-reversible alternative which we observe to often havefavourable convergence properties. We show how the Zig-Zag process can besimulated without discretisation error, and give conditions for the process tobe ergodic. Most importantly, we introduce a sub-sampling version of theZig-Zag process that is an example of an {\em exact approximate scheme}, i.e.the resulting approximate process still has the posterior as its stationarydistribution. Furthermore, if we use a control-variate idea to reduce thevariance of our unbiased estimator, then the Zig-Zag process can besuper-efficient: after an initial pre-processing step, essentially independentsamples from the posterior distribution are obtained at a computational costwhich does not depend on the size of the data.

The Zig-Zag Process and Super-Efficient Sampling for Bayesian Analysis of Big Data