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

While normalizing flows have led to significant advances in modelinghigh-dimensional continuous distributions, their applicability to discretedistributions remains unknown. In this paper, we show that flows can in fact beextended to discrete events---and under a simple change-of-variables formulanot requiring log-determinant-Jacobian computations. Discrete flows havenumerous applications. We consider two flow architectures: discreteautoregressive flows that enable bidirectionality, allowing, for example,tokens in text to depend on both left-to-right and right-to-left contexts in anexact language model; and discrete bipartite flows that enable efficientnon-autoregressive generation as in RealNVP. Empirically, we find that discreteautoregressive flows outperform autoregressive baselines on synthetic discretedistributions, an addition task, and Potts models; and bipartite flows canobtain competitive performance with autoregressive baselines on character-levellanguage modeling for Penn Tree Bank and text8.

Discrete Flows: Invertible Generative Models of Discrete Data