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

Knowledge distillation has been widely used to compress existing deeplearning models while preserving the performance on a wide range ofapplications. In the specific context of Automatic Speech Recognition (ASR),distillation from ensembles of acoustic models has recently shown promisingresults in increasing recognition performance. In this paper, we propose anextension of multi-teacher distillation methods to joint CTC-attentionend-to-end ASR systems. We also introduce three novel distillation strategies.The core intuition behind them is to integrate the error rate metric to theteacher selection rather than solely focusing on the observed losses. In thisway, we directly distill and optimize the student toward the relevant metricfor speech recognition. We evaluate these strategies under a selection oftraining procedures on different datasets (TIMIT, Librispeech, Common Voice)and various languages (English, French, Italian). In particular,state-of-the-art error rates are reported on the Common Voice French, Italianand TIMIT datasets.

Distilling Knowledge from Ensembles of Acoustic Models for Joint CTC-Attention End-to-End Speech Recognition