Research Papers 论文研究 1d ago Updated 20h ago 更新于 20小时前 45

Hear2Act: Benchmarking When Prosody Should Change What an Assistant Does Hear2Act:基准测试语音韵律何时应改变助手的行为

Hear2Act introduces a unified evaluation protocol with 480 persona-grounded scenarios to test whether prosodic cues in speech alter downstream task-oriented dialogue decisions Audio-capable LLMs show minimal improvement (14.6% to 15.3% optimal-solution rate) when relying solely on prosody-mediated feedback without explicit lexical signals Explicitly inferring concern status from audio and representing it in text before action selection boosts performance to 39.6%, approaching the 40.7% ceiling w 韵律线索(prosody)可传达任务相关信息,即使词汇内容不变也能改变对话轨迹和结果 提出Hear2Act统一评估协议,包含480个基于角色设定的场景,用于测试韵律证据如何影响下游决策 音频LLM仅靠音频输入时,最优解决方案率仅从14.6%微增至15.3%,韵律信息难以直接转化为行动 当模型将韵律推断的关注状态显式表示为文本后再用于行动选择,成功率提升至39.6%(接近40.7%的ground-truth水平) 韵律价值主要体现在词汇证据不足的场景,且需要显式中间表示才能可靠传递到决策环节

58
Hot 热度
72
Quality 质量
65
Impact 影响力

Analysis 深度分析

TL;DR

  • Hear2Act introduces a unified evaluation protocol with 480 persona-grounded scenarios to test whether prosodic cues in speech alter downstream task-oriented dialogue decisions
  • Audio-capable LLMs show minimal improvement (14.6% to 15.3% optimal-solution rate) when relying solely on prosody-mediated feedback without explicit lexical signals
  • Explicitly inferring concern status from audio and representing it in text before action selection boosts performance to 39.6%, approaching the 40.7% ceiling with ground-truth state
  • Prosody matters primarily when lexical evidence is insufficient; raw audio input alone is unreliable for driving assistant actions without an intermediate textual representation

Why It Matters

This benchmark addresses a critical gap in evaluating spoken AI assistants by isolating the impact of prosody on decision-making, rather than treating perception and action as separate problems. For practitioners building voice-based agents, it reveals that simply adding audio input is insufficient—structured reasoning over prosodic signals is essential for reliable task performance.

Technical Details

  • Hear2Act Benchmark: 480 persona-grounded scenarios with hidden user concerns and objectively verifiable outcomes; each scenario varies whether concerns are conveyed explicitly in words or primarily through prosody
  • Three evaluation conditions: transcript-only access, audio access, and concern-state access, enabling direct comparison of how prosodic information influences downstream decisions
  • Experimental setup: Two audio-capable LLMs evaluated under prosody-mediated feedback versus explicit concern inference with textual representation
  • Key metric: Optimal-solution rate, measuring whether the assistant selects the correct next action given the user's hidden concern
  • Findings: Prosody alone yields negligible gains (14.6% → 15.3%), but explicit inference from audio to text representation recovers most of the performance gap (39.6% vs. 40.7% ground truth)

Industry Insight

  • Voice assistant developers should invest in intermediate reasoning layers that explicitly extract and represent prosodic information in text before action selection, rather than relying on end-to-end audio-to-action pipelines
  • Benchmark design should decouple prosodic perception from task execution to accurately diagnose where failures occur in spoken dialogue systems
  • The results suggest a practical architecture: audio encoder → prosodic feature extraction → textual concern representation → decision module, which can bridge the gap between perception and action in spoken AI assistants

TL;DR

  • 韵律线索(prosody)可传达任务相关信息,即使词汇内容不变也能改变对话轨迹和结果
  • 提出Hear2Act统一评估协议,包含480个基于角色设定的场景,用于测试韵律证据如何影响下游决策
  • 音频LLM仅靠音频输入时,最优解决方案率仅从14.6%微增至15.3%,韵律信息难以直接转化为行动
  • 当模型将韵律推断的关注状态显式表示为文本后再用于行动选择,成功率提升至39.6%(接近40.7%的ground-truth水平)
  • 韵律价值主要体现在词汇证据不足的场景,且需要显式中间表示才能可靠传递到决策环节

为什么值得看

本文为语音助手系统提供了首个系统性基准,揭示了当前音频LLM在"听"与"做"之间的关键能力缺口。对AI从业者而言,研究结果直接指导了多模态对话系统的架构设计方向——单纯端到端音频输入不足以充分利用韵律信息,需要显式的中间表示层。

技术解析

  • Hear2Act基准设计:480个persona-grounded场景,每个场景保持任务和用户需求固定,仅改变关注点(concern)的表达方式(显式词汇 vs. 主要依赖韵律),并设置三种评估条件:仅文本转录、音频输入、以及关注状态已知。
  • 评估协议创新:首次将韵律感知、响应适当性和任务导向决策纳入统一框架,解决了现有基准孤立评估各能力、无法验证韵律证据是否真正改变下游决策的问题。
  • 关键实验发现:在Prosody-mediated feedback条件下,添加音频对决策影响极小(14.6%→15.3%);但当模型先推断关注状态并以文本形式显式表示,再用于行动选择时,成功率跃升至39.6%,接近使用ground-truth状态的40.7%。
  • 词汇 vs. 韵律的交互效应:在Explicit lexical feedback条件下(关注点已在话语中明确提及),音频与纯文本的性能差异基本消失,说明韵律的价值主要体现在词汇信息不足时的补充作用。

行业启示

  • 架构设计建议:语音助手系统不应依赖端到端的音频→行动映射,而应引入显式的"韵律理解→状态表示→决策"中间层,将隐式声学信息转化为可操作的文本表征。
  • 评估标准升级:行业需要超越单一模态的基准测试,建立能验证跨模态信息传递有效性的统一评估协议,Hear2Act为此提供了可复用的方法论框架。
  • 资源分配策略:在词汇信息充足的场景下,音频输入的边际价值有限;研发资源应优先投入韵律-语义融合机制和中间表示学习,而非单纯扩大音频训练数据。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

Speech 语音 Benchmark 基准测试 Evaluation 评测 Conversational AI 对话系统 Dataset 数据集