Research

Paper

AI LLM March 25, 2026

Semantic-Aware Interruption Detection in Spoken Dialogue Systems: Benchmark, Metric, and Model

Authors

Kangxiang Xia, Bingshen Mu, Xian Shi, Jin Xu, Lei Xie

Abstract

Achieving natural full-duplex interaction in spoken dialogue systems (SDS) remains a challenge due to the difficulty of accurately detecting user interruptions. Current solutions are polarized between "trigger-happy" VAD-based methods that misinterpret backchannels and robust end-to-end models that exhibit unacceptable response delays. Moreover, the absence of real-world benchmarks and holistic metrics hinders progress in the field. This paper presents a comprehensive frame-work to overcome these limitations. We first introduce SID-Bench, the first benchmark for semantic-aware interruption detection built entirely from real-world human dialogues. To provide a rigorous assessment of the responsiveness-robustness trade-off, we propose the Average Penalty Time (APT) metric, which assigns a temporal cost to both false alarms and late responses. Building on this framework, we design an LLM-based detection model optimized through a novel training paradigm to capture subtle semantic cues of intent. Experimental results show that our model significantly outperforms mainstream baselines, achieving a nearly threefold reduction in APT. By successfully resolving the long-standing tension between speed and stability, our work establishes a new state-of-the-art for intelligent interruption handling in SDS. To facilitate future research, SID-Bench and the associated code are available at: https://github.com/xkx-hub/SID-bench.

Metadata

arXiv ID: 2603.24144
Provider: ARXIV
Primary Category: cs.SD
Published: 2026-03-25
Fetched: 2026-03-26 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.24144v1</id>\n    <title>Semantic-Aware Interruption Detection in Spoken Dialogue Systems: Benchmark, Metric, and Model</title>\n    <updated>2026-03-25T10:11:33Z</updated>\n    <link href='https://arxiv.org/abs/2603.24144v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.24144v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Achieving natural full-duplex interaction in spoken dialogue systems (SDS) remains a challenge due to the difficulty of accurately detecting user interruptions. Current solutions are polarized between \"trigger-happy\" VAD-based methods that misinterpret backchannels and robust end-to-end models that exhibit unacceptable response delays. Moreover, the absence of real-world benchmarks and holistic metrics hinders progress in the field. This paper presents a comprehensive frame-work to overcome these limitations. We first introduce SID-Bench, the first benchmark for semantic-aware interruption detection built entirely from real-world human dialogues. To provide a rigorous assessment of the responsiveness-robustness trade-off, we propose the Average Penalty Time (APT) metric, which assigns a temporal cost to both false alarms and late responses. Building on this framework, we design an LLM-based detection model optimized through a novel training paradigm to capture subtle semantic cues of intent. Experimental results show that our model significantly outperforms mainstream baselines, achieving a nearly threefold reduction in APT. By successfully resolving the long-standing tension between speed and stability, our work establishes a new state-of-the-art for intelligent interruption handling in SDS. To facilitate future research, SID-Bench and the associated code are available at: https://github.com/xkx-hub/SID-bench.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.SD'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='eess.AS'/>\n    <published>2026-03-25T10:11:33Z</published>\n    <arxiv:comment>Accepted by ICME 2026</arxiv:comment>\n    <arxiv:primary_category term='cs.SD'/>\n    <author>\n      <name>Kangxiang Xia</name>\n    </author>\n    <author>\n      <name>Bingshen Mu</name>\n    </author>\n    <author>\n      <name>Xian Shi</name>\n    </author>\n    <author>\n      <name>Jin Xu</name>\n    </author>\n    <author>\n      <name>Lei Xie</name>\n    </author>\n  </entry>"
}