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/

Zendy Plus

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

Zendy Tools

Zendy Open

Recently, Mixture of Experts (MoE) based Transformer has shown promisingresults in many domains. This is largely due to the following advantages ofthis architecture: firstly, MoE based Transformer can increase model capacitywithout computational cost increasing both at training and inference time.Besides, MoE based Transformer is a dynamic network which can adapt to thevarying complexity of input instances in realworld applications. In this work,we explore the MoE based model for speech recognition, named SpeechMoE. Tofurther control the sparsity of router activation and improve the diversity ofgate values, we propose a sparsity L1 loss and a mean importance lossrespectively. In addition, a new router architecture is used in SpeechMoE whichcan simultaneously utilize the information from a shared embedding network andthe hierarchical representation of different MoE layers. Experimental resultsshow that SpeechMoE can achieve lower character error rate (CER) withcomparable computation cost than traditional static networks, providing7.0%-23.0% relative CER improvements on four evaluation datasets.

SpeechMoE: Scaling to Large Acoustic Models with Dynamic Routing Mixture of Experts