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

In dysarthric speech recognition, data scarcity and the vast diversitybetween dysarthric speakers pose significant challenges. While finetuning hasbeen a popular solution, it can lead to overfitting and low parameterefficiency. Adapter modules offer a better solution, with their small size andeasy applicability. Additionally, Adapter Fusion can facilitate knowledgetransfer from multiple learned adapters, but may employ more parameters. Inthis work, we apply Adapter Fusion for target speaker adaptation and speechrecognition, achieving acceptable accuracy with significantly fewerspeaker-specific trainable parameters than classical finetuning methods. Wefurther improve the parameter efficiency of the fusion layer by reducing thesize of query and key layers and using Householder transformation toreparameterize the value linear layer. Our proposed fusion layer achievescomparable recognition results to the original method with only one third ofthe parameters.

Parameter-efficient Dysarthric Speech Recognition Using Adapter Fusion and Householder Transformation