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

As speech-enabled devices such as smartphones and smart speakers becomeincreasingly ubiquitous, there is growing interest in building automatic speechrecognition (ASR) systems that can run directly on-device; end-to-end (E2E)speech recognition models such as recurrent neural network transducers andtheir variants have recently emerged as prime candidates for this task. Apartfrom being accurate and compact, such systems need to decode speech with lowuser-perceived latency (UPL), producing words as soon as they are spoken. Thiswork examines the impact of various techniques - model architectures, trainingcriteria, decoding hyperparameters, and endpointer parameters - on UPL. Ouranalyses suggest that measures of model size (parameters, input chunk sizes),or measures of computation (e.g., FLOPS, RTF) that reflect the model's abilityto process input frames are not always strongly correlated with observed UPL.Thus, conventional algorithmic latency measurements might be inadequate inaccurately capturing latency observed when models are deployed on embeddeddevices. Instead, we find that factors affecting token emission latency, andendpointing behavior have a larger impact on UPL. We achieve the best trade-offbetween latency and word error rate when performing ASR jointly withendpointing, while utilizing the recently proposed alignment regularizationmechanism.

Dissecting User-Perceived Latency of On-Device E2E Speech Recognition