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

Token-level serialized output training (t-SOT) was recently proposed toaddress the challenge of streaming multi-talker automatic speech recognition(ASR). T-SOT effectively handles overlapped speech by representing multi-talkertranscriptions as a single token stream with $\langle \text{cc}\rangle$ symbolsinterspersed. However, the use of a naive neural transducer architecturesignificantly constrained its applicability for text-only adaptation. Toovercome this limitation, we propose a novel t-SOT model structure thatincorporates the idea of factorized neural transducers (FNT). The proposedmethod separates a language model (LM) from the transducer's predictor andhandles the unnatural token order resulting from the use of $\langle\text{cc}\rangle$ symbols in t-SOT. We achieve this by maintaining multiplehidden states and introducing special handling of the $\langle\text{cc}\rangle$ tokens within the LM. The proposed t-SOT FNT model achievescomparable performance to the original t-SOT model while retaining the abilityto reduce word error rate (WER) on both single and multi-talker datasetsthrough text-only adaptation.

t-SOT FNT: Streaming Multi-talker ASR with Text-only Domain Adaptation Capability