all InfoSec news
Extracting Targeted Training Data from ASR Models, and How to Mitigate It. (arXiv:2204.08345v2 [cs.SD] UPDATED)
June 29, 2022, 1:20 a.m. | Ehsan Amid, Om Thakkar, Arun Narayanan, Rajiv Mathews, Françoise Beaufays
cs.CR updates on arXiv.org arxiv.org
Recent work has designed methods to demonstrate that model updates in ASR
training can leak potentially sensitive attributes of the utterances used in
computing the updates. In this work, we design the first method to demonstrate
information leakage about training data from trained ASR models. We design
Noise Masking, a fill-in-the-blank style method for extracting targeted parts
of training data from trained ASR models. We demonstrate the success of Noise
Masking by using it in four settings for extracting names …
More from arxiv.org / cs.CR updates on arXiv.org
Jobs in InfoSec / Cybersecurity
Sr. Cloud Security Engineer
@ BLOCKCHAINS | USA - Remote
Network Security (SDWAN: Velocloud) Infrastructure Lead
@ Sopra Steria | Noida, Uttar Pradesh, India
Senior Python Engineer, Cloud Security
@ Darktrace | Cambridge
Senior Security Consultant
@ Nokia | United States
Manager, Threat Operations
@ Ivanti | United States, Remote
Lead Cybersecurity Architect - Threat Modeling | AWS Cloud Security
@ JPMorgan Chase & Co. | Columbus, OH, United States