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

Probabilistic models based on Restricted Boltzmann Machines (RBMs) imply theevaluation of normalized Boltzmann factors, which in turn require from theevaluation of the partition function Z. The exact evaluation of Z, though,becomes a forbiddingly expensive task as the system size increases. This evenworsens when one considers most usual learning algorithms for RBMs, where theexact evaluation of the gradient of the log-likelihood of the empiricaldistribution of the data includes the computation of Z at each iteration. TheAnnealed Importance Sampling (AIS) method provides a tool to stochasticallyestimate the partition function of the system. So far, the standard use of theAIS algorithm in the Machine Learning context has been done using a largenumber of Monte Carlo steps. In this work we show that this may not be requiredif a proper starting probability distribution is employed as the initializationof the AIS algorithm. We analyze the performance of AIS in both small- andlarge-sized problems, and show that in both cases a good estimation of Z can beobtained with little computational cost.

Efficient Evaluation of the Partition Function of RBMs with Annealed Importance Sampling