New method masks dialogue system response delays with short-response selection

Researchers have developed a novel real-time short-response selection task designed to improve the naturalness of spoken dialogue systems. This method uses contrastive learning to map dialogue context to pre-generated short responses, which are played to mask delays in main response generation and speech synthesis. Experiments indicate that this approach effectively selects contextually appropriate short responses, outperforming random and LLM-based baselines while maintaining turn-taking prediction performance with minimal overhead. AI

IMPACT Improves naturalness in spoken dialogue systems by masking generation delays.

RANK_REASON The item is an academic paper detailing a new method for spoken dialogue systems. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• Gotit.pub
• Hugging Face
• Influence Flower
• Litmaps
• ScienceCast
• scite Smart Citations

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →