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

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 desi...

Full description

Saved in:
Bibliographic Details
Main Authors: Dang, Trung, Thakkar, Om, Ramaswamy, Swaroop, Mathews, Rajiv, Chin, Peter, Beaufays, Francoise
Format: Conference Proceeding
Language:English
Subjects:
ASR
Computational modeling
Conferences
Design methodology
distributed training
Impedance matching
Optimization methods
privacy
Stochastic processes
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary: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 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).
ISSN:2379-190X
DOI:10.1109/ICASSP43922.2022.9746443