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

The representation learning of speech, without textual resources, is an areaof significant interest for many low resource speech applications. In thispaper, we describe an approach to self-supervised representation learning fromraw audio using a hidden unit clustering (HUC) framework. The input to themodel consists of audio samples that are windowed and processed with 1-Dconvolutional layers. The learned "time-frequency" representations from theconvolutional neural network (CNN) module are further processed with long shortterm memory (LSTM) layers which generate a contextual vector representation forevery windowed segment. The HUC framework, allowing the categorization of therepresentations into a small number of phoneme-like units, is used to train themodel for learning semantically rich speech representations. The targetsconsist of phoneme-like pseudo labels for each audio segment and these aregenerated with an iterative k-means algorithm. We explore techniques thatimprove the speaker invariance of the learned representations and illustratethe effectiveness of the proposed approach on two settings, i) completelyunsupervised speech applications on the sub-tasks described as part of theZeroSpeech 2021 challenge and ii) semi-supervised automatic speech recognition(ASR) applications on the TIMIT dataset and on the GramVaani challenge Hindidataset. In these experiments, we achieve state-of-art results for variousZeroSpeech tasks. Further, on the ASR experiments, the HUC representations areshown to improve significantly over other established benchmarks based onWav2vec, HuBERT and Best-RQ.

Representation Learning With Hidden Unit Clustering For Low Resource Speech Applications