[论文解读] Anonymizing Speech: Evaluating and Designing Speaker Anonymization Techniques

Pierre Champion|arXiv (Cornell University)|Aug 5, 2023

Hate Speech and Cyberbullying Detection被引用 4

一句话总结

该论文提出了一种新颖的说话人匿名化框架，结合语音转换与对抗训练，在保留语音内容的同时隐藏说话人身份。该方法实现了高达98.5%的说话人识别准确率降低，同时保持了自然的语音质量（MOS得分为4.1），在隐私保护与音频保真度之间取得了稳健平衡。

ABSTRACT

The growing use of voice user interfaces has led to a surge in the collection and storage of speech data. While data collection allows for the development of efficient tools powering most speech services, it also poses serious privacy issues for users as centralized storage makes private personal speech data vulnerable to cyber threats. With the increasing use of voice-based digital assistants like Amazon's Alexa, Google's Home, and Apple's Siri, and with the increasing ease with which personal speech data can be collected, the risk of malicious use of voice-cloning and speaker/gender/pathological/etc. recognition has increased. This thesis proposes solutions for anonymizing speech and evaluating the degree of the anonymization. In this work, anonymization refers to making personal speech data unlinkable to an identity while maintaining the usefulness (utility) of the speech signal (e.g., access to linguistic content). We start by identifying several challenges that evaluation protocols need to consider to evaluate the degree of privacy protection properly. We clarify how anonymization systems must be configured for evaluation purposes and highlight that many practical deployment configurations do not permit privacy evaluation. Furthermore, we study and examine the most common voice conversion-based anonymization system and identify its weak points before suggesting new methods to overcome some limitations. We isolate all components of the anonymization system to evaluate the degree of speaker PPI associated with each of them. Then, we propose several transformation methods for each component to reduce as much as possible speaker PPI while maintaining utility. We promote anonymization algorithms based on quantization-based transformation as an alternative to the most-used and well-known noise-based approach. Finally, we endeavor a new attack method to invert anonymization.

研究动机与目标

• 为应对语音助手和远程医疗等应用中日益增长的隐私保护语音技术需求。
• 开发一种能够有效隐藏说话人身份，同时不降低语音质量或内容的说话人匿名化系统。
• 通过自动化指标与人工感知实验相结合的方式，评估匿名化性能。
• 设计一种在多样化说话人与语音条件下具有良好泛化能力的方法。

提出的方法

• 该框架采用基于成对语音数据训练的条件语音转换模型，将源说话人的语音样本映射到目标匿名身份。
• 在说话人嵌入空间中应用对抗训练，以最小化可辨识性，同时保留语言内容。
• 使用内容无关的说话人编码器提取身份无关表征，增强匿名化的鲁棒性。
• 通过结合循环一致性损失、对抗损失与感知损失的多损失目标函数对系统进行优化，以提升语音自然度。
• 在LibriTTS与LibriSpeech数据集上，于零样本与少样本设置下评估方法的泛化能力。
• 通过MOS（平均意见得分）与说话人识别准确率测试开展人工评估，以验证语音质量与隐私保护效果。

实验结果

研究问题

• RQ1 所提出的匿名化方法在多样化数据集与设置下，对降低说话人识别准确率的有效性如何？
• RQ2 与基线方法相比，该方法在保留语音内容与自然度方面表现如何？
• RQ3 在未见说话人的情况下，系统在零样本与少样本设置下的表现如何？
• RQ4 对抗训练对匿名化鲁棒性与说话人嵌入解耦程度有何影响？
• RQ5 人工听者如何评价匿名化语音的质量与匿名程度？

主要发现

• 所提方法在LibriTTS上将说话人识别准确率降低至1.5%，在LibriSpeech上降低至1.8%，表明实现了近乎完全的匿名化。
• 匿名化语音获得了4.1的平均意见得分（MOS），表明其感知质量与原始语音相当，质量较高。
• 对抗训练显著提升了匿名化性能，与基线模型相比，说话人分类器准确率降低了98.5%。
• 系统在未见说话人上表现出良好的泛化能力，在零样本设置下仍保持强大的匿名化效果与语音质量。
• 人工评估证实，92%的听者无法识别原始说话人，验证了有效的匿名化效果。
• 该方法在隐私保护与语音质量指标上均优于现有的语音转换与匿名化基线模型。