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

Large-scale conversational assistants like Alexa, Siri, Cortana and GoogleAssistant process every utterance using multiple models for domain, intent andnamed entity recognition. Given the decoupled nature of model development andlarge traffic volumes, it is extremely difficult to identify utterancesprocessed erroneously by such systems. We address this challenge to detectdomain classification errors using offline Transformer models. We combineutterance encodings from a RoBERTa model with the Nbest hypothesis produced bythe production system. We then fine-tune end-to-end in a multitask settingusing a small dataset of humanannotated utterances with domain classificationerrors. We tested our approach for detecting misclassifications from one domainthat accounts for <0.5% of the traffic in a large-scale conversational AIsystem. Our approach achieves an F1 score of 30% outperforming a bi- LSTMbaseline by 16.9% and a standalone RoBERTa model by 4.8%. We improve thisfurther by 2.2% to 32.2% by ensembling multiple models.

Error Detection in Large-Scale Natural Language Understanding Systems Using Transformer Models