Title: FeatureSense: Protecting Speaker Attributes in Always-On Audio Sensing System Show Chinese title

Title: FeatureSense：保护始终在线音频传感系统中的说话者属性

Abstract: Audio is a rich sensing modality that is useful for a variety of human activity recognition tasks. However, the ubiquitous nature of smartphones and smart speakers with always-on microphones has led to numerous privacy concerns and a lack of trust in deploying these audio-based sensing systems. This paper addresses this critical challenge of preserving user privacy when using audio for sensing applications while maintaining utility. While prior work focuses primarily on protecting recoverable speech content, we show that sensitive speaker-specific attributes such as age and gender can still be inferred after masking speech and propose a comprehensive privacy evaluation framework to assess this speaker attribute leakage. We design and implement FeatureSense, an open-source library that provides a set of generalizable privacy-aware audio features that can be used for wide range of sensing applications. We present an adaptive task-specific feature selection algorithm that optimizes the privacy-utility-cost trade-off based on the application requirements. Through our extensive evaluation, we demonstrate the high utility of FeatureSense across a diverse set of sensing tasks. Our system outperforms existing privacy techniques by 60.6% in preserving user-specific privacy. This work provides a foundational framework for ensuring trust in audio sensing by enabling effective privacy-aware audio classification systems. Show Chinese abstract

Abstract: 音频是一种丰富的感知模态，在各种人类活动识别任务中非常有用。然而，智能手机和智能音箱等设备的无处不在，加上它们始终开启的麦克风功能，引发了众多隐私问题，并导致人们对部署这些基于音频的感知系统缺乏信任。本文解决了在使用音频进行感知应用时保护用户隐私这一关键挑战，同时保持实用性。虽然之前的研究主要集中在保护可恢复的语音内容上，但我们展示了即使在屏蔽语音后，敏感的说话者特定属性（如年龄和性别）仍然可以被推断出来，并提出了一种全面的隐私评估框架来评估这种说话者属性泄漏。我们设计并实现了FeatureSense，这是一个开源库，提供了一组通用的隐私感知音频特征，可用于广泛的感知应用。我们提出了一个自适应的任务特定特征选择算法，根据应用需求优化隐私-效用-成本权衡。通过我们的广泛评估，我们证明了FeatureSense在各种感知任务中的高实用性。在保护特定用户隐私方面，我们的系统比现有的隐私技术高出60.6%。这项工作为确保音频感知的信任提供了一个基础框架，使有效的隐私感知音频分类系统成为可能。