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

The effective exploitation of richer contextual information in languagemodels (LMs) is a long-standing research problem for automatic speechrecognition (ASR). A cross-utterance LM (CULM) is proposed in this paper, whichaugments the input to a standard long short-term memory (LSTM) LM with acontext vector derived from past and future utterances using an extractionnetwork. The extraction network uses another LSTM to encode surroundingutterances into vectors which are integrated into a context vector using eithera projection of LSTM final hidden states, or a multi-head self-attentive layer.In addition, an acoustic error sampling technique is proposed to reduce themismatch between training and test-time. This is achieved by consideringpossible ASR errors into the model training procedure, and can thereforeimprove the word error rate (WER). Experiments performed on both AMI andSwitchboard datasets show that CULMs outperform the LSTM LM baseline WER. Inparticular, the CULM with a self-attentive layer-based extraction network andacoustic error sampling achieves 0.6% absolute WER reduction on AMI, 0.3% WERreduction on the Switchboard part and 0.9% WER reduction on the Callhome partof Eval2000 test set over the respective baselines.

Cross-Utterance Language Models with Acoustic Error Sampling