AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

SpokenWOZ: A Large-Scale Speech-Text Benchmark for Spoken Task-Oriented Dialogue Agents

Shuzheng Si, Wentao Ma, Haoyu Gao, Yuchuan Wu, Ting-En Lin, Yinpei Dai, Hangyu Li, Rui Yan, Fei Huang, Yongbin Li

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: Speech and audio language models · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add aligned WavLM speech embeddings to SPACESPACE · SpokenWOZ DSTJGA: 22.73 -> 25.65Improved
Add WavLM speech embeddings without alignmentSPACE · SpokenWOZ DSTJGA: 22.73 -> 24.09Improved
Align words with audio pieces (WavLMalign vs WavLM)SPACE+WavLM · SpokenWOZ DSTJGA: 24.09 -> 25.65Improved
Add WavLM fusion to SPACE+TripPySPACE+TripPy · SpokenWOZ DSTJGA: 16.24 -> 18.71Improved
Replace BERT encoder by SPACE in TripPyBERT+TripPy · SpokenWOZ DSTJGA: 14.78 -> 16.24Improved
Generative (UBAR) instead of extractive (TripPy) DSTBERT+TripPy · SpokenWOZ DSTJGA: 14.78 -> 20.54Improved
Exclude cross-turn slots from evaluationSPACE · SpokenWOZ DSTJGA: 22.73 -> 26.99Improved
Add aligned speech to SPACE in policy optimizationSPACE · SpokenWOZ policy optimizationComb: 85.52 -> 88.01Improved
Add aligned speech to SPACE in end-to-end modelingSPACE · SpokenWOZ end-to-endComb: 79.84 -> 82.32Improved
Align speech vs unaligned WavLM in end-to-end modelingSPACE+WavLM · SpokenWOZ end-to-endComb: 80.71 -> 82.32Improved
Add WavLM without alignment to SPACE, policy BLEUSPACE · SpokenWOZ policy optimizationBLEU: 18.72 -> 18.54No clear effect
Evaluate SPACE on spoken (SpokenWOZ) instead of written MultiWOZSPACE · DSTJGA: 57.5 -> 22.73Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy