MultiPartyTurnBench：多轮口语对话中的话语权控制基准测试
MultiPartyTurnBench: Benchmarking Floor Control in Multi-Party Spoken Dialogue
AI总结：
该研究推出多场景话语权控制基准MultiPartyTurnBench，评估话轮结束检测与中断检测任务，发现现有系统在中断检测上表现接近随机，选择性话语权控制仍是多场景语音交互的未解决挑战。
摘要：
全双工语音系统必须决定何时接入、保持或放弃对话话语权，即便存在多位说话人重叠或未获取话语权的反馈（backchannel）情况。现有评估多聚焦于双人对话的话轮转换或端到端智能体行为，缺乏针对多场景下真实话语权转换的统一、可控评估，且未区分易混淆的非事件。我们推出MultiPartyTurnBench（MPTB），这是一个包含240个合成生成、通道分离的对话基准，覆盖2至5位说话人和6种交互类型。MPTB评估两项互补任务：话轮结束（EOT）检测，用于区分真实话轮完成与临时保持；中断（INT）检测，用于区分成功获取话语权与未获取话语权的语音。精确的事件条件注释支持使用假阳性约束召回率和带符号延迟进行因果评估。在声学、语义和话轮投影基线中，我们发现显著的不对称性：学习系统在4.3%至11.8%的假阳性率下，EOT召回率达到82.5%至92.6%，而INT检测则存在错过获取或过度误激活的问题。这种不对称性延伸至5个全双工智能体，它们在话轮接入上达到86.1%至95.8%的平衡准确率，但在选择性中断处理上表现接近随机（46.9%至54.7%）。Gemini-3.8-Live对97.7%的反馈（backchannel）弃权（不执行），而MiniCPM-o-4.5仅对1.4%的真实中断弃权（不执行）。这些发现表明，选择性话语权控制是可靠多场景语音交互中一个独特且未解决的挑战。
英文摘要：
Full-duplex speech systems must decide when to enter, hold, or relinquish the conversational floor, even as multiple speakers overlap or backchannel without taking it. Existing evaluations study dyadic turn-taking or end-to-end agent behavior, but lack a unified, controlled evaluation of genuine floor transitions against confusable non-events in multi-party settings. We introduce MultiPartyTurnBench (MPTB), a benchmark of 240 synthetically generated, channel-separated conversations spanning two to five speakers and six interaction types. MPTB evaluates two complementary tasks: End-of-Turn (EOT) detection distinguishes genuine turn completion from temporary holds, while Interruption (INT) detection distinguishes successful floor takeovers from non-floor-taking speech. Exact event-conditioned annotations enable causal evaluation using false-positive-constrained recall and signed latency. Across acoustic, semantic, and turn-projection baselines, we find a stark asymmetry: learned systems achieve 82.5-92.6% EOT recall at 4.3-11.8% false-positive rates, while INT detection suffers from missed takeovers or excessive false activations. This asymmetry extends to five full-duplex agents which achieve 86.1-95.8% balanced accuracy on turn entry, yet perform near chance (46.9-54.7%) on selective interruption handling. Gemini-3.8-Live yields to 97.7% of backchannels, whereas MiniCPM-o-4.5 yields to only 1.4% of genuine interruptions. These findings establish selective floor control as a distinct, unresolved challenge for reliable multi-party voice interaction.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
