WaveCommitments：隐私可验证的语音智能体市场

AI 摘要

论文提出 WaveCommitments，一个开源的端到端设计，将可信执行环境（TEE）内的 Data Shapley 保密资产估值与多元多项式承诺结合，用于语音智能体市场。它支持选择性打开和防篡改溯源，证明交付资产与估值一致，同时隐藏其余信息。在三个数据集和 Whisper 等编码器上的实验显示，最佳基准最高加速 20 倍、证明体积缩小 39 倍，并揭示了说话人价值随任务和编码器变化的市场价格。

正文节选

Baser Kale Vishwanath Chinchali WaveCommitments: Private and Verifiable Speech-Agent Marketplaces Abstract Modern speech systems are evolving into ecosystems of autonomous speech agents that speak together, negotiate, and trade assets such as embeddings, spectrograms, or raw data. In such markets, sellers cannot reveal assets for evaluation due to privacy constraints, and buyers cannot price what they cannot test. Thus, we open-source WaveCommitments, an end-to-end design that pairs confidential

发布时间：2026-10-21 12:00

抓取时间：2026-10-21 12:26

来源机构：arXiv