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

Self-supervised learning (SSL) for rich speech representations has achievedempirical success in low-resource Automatic Speech Recognition (ASR) and otherspeech processing tasks, which can mitigate the necessity of a large amount oftranscribed speech and thus has driven a growing demand for on-device ASR andother speech processing. However, advanced speech SSL models have becomeincreasingly large, which contradicts the limited on-device resources. This gapcould be more severe in multilingual/multitask scenarios requiringsimultaneously recognizing multiple languages or executing multiple speechprocessing tasks. Additionally, strongly overparameterized speech SSL modelstend to suffer from overfitting when being finetuned on low-resource speechcorpus. This work aims to enhance the practical usage of speech SSL modelstowards a win-win in both enhanced efficiency and alleviated overfitting viaour proposed S$^3$-Router framework, which for the first time discovers thatsimply discarding no more than 10\% of model weights via only finetuning modelconnections of speech SSL models can achieve better accuracy over standardweight finetuning on downstream speech processing tasks. More importantly,S$^3$-Router can serve as an all-in-one technique to enable (1) a newfinetuning scheme, (2) an efficient multilingual/multitask solution, (3) astate-of-the-art ASR pruning technique, and (4) a new tool to quantitativelyanalyze the learned speech representation. We believe S$^3$-Router has provideda new perspective for practical deployment of speech SSL models. Our codes areavailable at: https://github.com/GATECH-EIC/S3-Router.

Losses Can Be Blessings: Routing Self-Supervised Speech Representations Towards Efficient Multilingual and Multitask Speech Processing