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

Generative models have long been the dominant approach for speechrecognition. The success of these models however relies on the use ofsophisticated recipes and complicated machinery that is not easily accessibleto non-practitioners. Recent innovations in Deep Learning have given rise to analternative - discriminative models called Sequence-to-Sequence models, thatcan almost match the accuracy of state of the art generative models. Whilethese models are easy to train as they can be trained end-to-end in a singlestep, they have a practical limitation that they can only be used for offlinerecognition. This is because the models require that the entirety of the inputsequence be available at the beginning of inference, an assumption that is notvalid for instantaneous speech recognition. To address this problem, onlinesequence-to-sequence models were recently introduced. These models are able tostart producing outputs as data arrives, and the model feels confident enoughto output partial transcripts. These models, like sequence-to-sequence arecausal - the output produced by the model until any time, $t$, affects thefeatures that are computed subsequently. This makes the model inherently morepowerful than generative models that are unable to change features that arecomputed from the data. This paper highlights two main contributions - animprovement to online sequence-to-sequence model training, and its applicationto noisy settings with mixed speech from two speakers.

An online sequence-to-sequence model for noisy speech recognition