arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.07815cs.CLcs.CRcs.LG
收藏

分布式ASR训练中的说话人身份识别方法及其对抗

A Method to Reveal Speaker Identity in Distributed ASR Training, and How to Counter It

Boston University(波士顿大学);Google(谷歌)

Trung Dang, Om Thakkar, Swaroop Ramaswamy, Rajiv Mathews, Peter Chin, Françoise Beaufays

摘要:

端到端自动语音识别(ASR)模型通常使用随机梯度下降(SGD)等优化方法在口语上进行训练。在像联邦学习这样的分布式环境中,模型训练需要通过网络传输梯度。在这项工作中,我们设计了第一种方法,用于揭示一个训练话语的扬声器的身份,只访问一个梯度。我们提出了Hessian-Free的特征匹配,一种输入重建技术,无需损失函数的二阶导数(在以前的作品中需要),这可能是昂贵的计算。我们使用DeepSpeech模型架构展示了我们方法的有效性,证明在LibriSpeech数据集上可以以34%的前1准确度(51%的前5准确度)揭示说话者的身份。此外,我们研究了两个著名的技术,差分私人SGD和辍学,我们的方法的成功的效果。我们发现,0.2的辍学率可以降低扬声器身份的准确性为0%的前1(0.5%的前5)。

英文摘要:

End-to-end Automatic Speech Recognition (ASR) models are commonly trained over spoken utterances using optimization methods like Stochastic Gradient Descent (SGD). In distributed settings like Federated Learning, model training requires transmission of gradients over a network. In this work, we design the first method for revealing the identity of the speaker of a training utterance with access only to a gradient. We propose Hessian-Free Gradients Matching, an input reconstruction technique that operates without second derivatives of the loss function (required in prior works), which can be expensive to compute. We show the effectiveness of our method using the DeepSpeech model architecture, demonstrating that it is possible to reveal the speaker's identity with 34% top-1 accuracy (51% top-5 accuracy) on the LibriSpeech dataset. Further, we study the effect of two well-known techniques, Differentially Private SGD and Dropout, on the success of our method. We show that a dropout rate of 0.2 can reduce the speaker identity accuracy to 0% top-1 (0.5% top-5).

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。

↑