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

Open-vocabulary keyword spotting (KWS), which allows users to customizekeywords, has attracted increasingly more interest. However, existing methodsbased on acoustic models and post-processing train the acoustic model with ASRtraining criteria to model all phonemes, making the acoustic modelunder-optimized for the KWS task. To solve this problem, we propose a novelunified two-pass open-vocabulary KWS (U2-KWS) framework inspired by thetwo-pass ASR model U2. Specifically, we employ the CTC branch as the firststage model to detect potential keyword candidates and the decoder branch asthe second stage model to validate candidates. In order to enhance anycustomized keywords, we redesign the U2 training procedure for U2-KWS and addkeyword information by audio and text cross-attention into both branches. Weperform experiments on our internal dataset and Aishell-1. The results showthat U2-KWS can achieve a significant relative wake-up rate improvement of 41%compared to the traditional customized KWS systems when the false alarm rate isfixed to 0.5 times per hour.

U2-KWS: Unified Two-pass Open-vocabulary Keyword Spotting with Keyword Bias