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

Spoken Language Understanding (SLU), a core component of the task-orienteddialogue system, expects a shorter inference facing the impatience of humanusers. Existing work increases inference speed by designing non-autoregressivemodels for single-turn SLU tasks but fails to apply to multi-turn SLU inconfronting the dialogue history. The intuitive idea is to concatenate allhistorical utterances and utilize the non-autoregressive models directly.However, this approach seriously misses the salient historical information andsuffers from the uncoordinated-slot problems. To overcome those shortcomings,we propose a novel model for multi-turn SLU named Salient History Attentionwith Layer-Refined Transformer (SHA-LRT), which composes of an SHA module, aLayer-Refined Mechanism (LRM), and a Slot Label Generation (SLG) task. SHAcaptures salient historical information for the current dialogue from bothhistorical utterances and results via a well-designed history-attentionmechanism. LRM predicts preliminary SLU results from Transformer's middlestates and utilizes them to guide the final prediction, and SLG obtains thesequential dependency information for the non-autoregressive encoder.Experiments on public datasets indicate that our model significantly improvesmulti-turn SLU performance (17.5% on Overall) with accelerating (nearly 15times) the inference process over the state-of-the-art baseline as well aseffective on the single-turn SLU tasks.

Capture Salient Historical Information: A Fast and Accurate Non-Autoregressive Model for Multi-turn Spoken Language Understanding