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

Inference and learning for probabilistic generative networks is often verychallenging and typically prevents scalability to as large networks as used fordeep discriminative approaches. To obtain efficiently trainable, large-scaleand well performing generative networks for semi-supervised learning, we herecombine two recent developments: a neural network reformulation of hierarchicalPoisson mixtures (Neural Simpletrons), and a novel truncated variational EMapproach (TV-EM). TV-EM provides theoretical guarantees for learning ingenerative networks, and its application to Neural Simpletrons results inparticularly compact, yet approximately optimal, modifications of learningequations. If applied to standard benchmarks, we empirically find, thatlearning converges in fewer EM iterations, that the complexity per EM iterationis reduced, and that final likelihood values are higher on average. For thetask of classification on data sets with few labels, learning improvementsresult in consistently lower error rates if compared to applications withouttruncation. Experiments on the MNIST data set herein allow for comparison tostandard and state-of-the-art models in the semi-supervised setting. Furtherexperiments on the NIST SD19 data set show the scalability of the approach whena manifold of additional unlabeled data is available.

Truncated Variational EM for Semi-Supervised Neural Simpletrons