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

Automatic Speech Recognition (ASR) systems are known to exhibit difficultieswhen transcribing children's speech. This can mainly be attributed to theabsence of large children's speech corpora to train robust ASR models and theresulting domain mismatch when decoding children's speech with systems trainedon adult data. In this paper, we propose multiple enhancements to alleviatethese issues. First, we propose a data augmentation technique based on thesource-filter model of speech to close the domain gap between adult andchildren's speech. This enables us to leverage the data availability of adultspeech corpora by making these samples perceptually similar to children'sspeech. Second, using this augmentation strategy, we apply transfer learning ona Transformer model pre-trained on adult data. This model follows the recentlyintroduced XLS-R architecture, a wav2vec 2.0 model pre-trained on severalcross-lingual adult speech corpora to learn general and robust acousticframe-level representations. Adopting this model for the ASR task using adultdata augmented with the proposed source-filter warping strategy and a limitedamount of in-domain children's speech significantly outperforms previousstate-of-the-art results on the PF-STAR British English Children's Speechcorpus with a 4.86% WER on the official test set.

Transfer Learning for Robust Low-Resource Children's Speech ASR with Transformers and Source-Filter Warping