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

Self-Supervised Learning (SSL) has made great strides recently. SSL speechmodels achieve decent performance on a wide range of downstream tasks,suggesting that they extract different aspects of information from speech.However, how SSL models store various information in hidden representationswithout interfering is still poorly understood. Taking the recently successfulSSL model, HuBERT, as an example, we explore how the SSL model processes andstores speaker information in the representation. We found that HuBERT storesspeaker information in representations whose positions correspond to silencesin a waveform. There are several pieces of evidence. (1) We find that theutterances with more silent parts in the waveforms have better SpeakerIdentification (SID) accuracy. (2) If we use the whole utterances for SID, thesilence part always contributes more to the SID task. (3) If we only use therepresentation of a part of the utterance for SID, the silenced part has higheraccuracy than the other parts. Our findings not only contribute to a betterunderstanding of SSL models but also improve performance. By simply addingsilence to the original waveform, HuBERT improved its accuracy on SID by nearly2%.

Silence is Sweeter Than Speech: Self-Supervised Model Using Silence to Store Speaker Information