all InfoSec news
Reconstructing Training Data from Model Gradient, Provably. (arXiv:2212.03714v1 [cs.LG])
Dec. 8, 2022, 2:18 a.m. | Zihan Wang, Jason Lee, Qi Lei
cs.CR updates on arXiv.org arxiv.org
Understanding when and how much a model gradient leaks information about the
training sample is an important question in privacy. In this paper, we present
a surprising result: even without training or memorizing the data, we can fully
reconstruct the training samples from a single gradient query at a randomly
chosen parameter value. We prove the identifiability of the training data under
mild conditions: with shallow or deep neural networks and a wide range of
activation functions. We also present …
More from arxiv.org / cs.CR updates on arXiv.org
Jobs in InfoSec / Cybersecurity
SOC 2 Manager, Audit and Certification
@ Deloitte | US and CA Multiple Locations
Security Engineering Professional
@ Nokia | India
Cyber Intelligence Exercise Planner
@ Peraton | Fort Gordon, GA, United States
Technical Lead, HR Systems Security
@ Sun Life | Sun Life Wellesley
SecOps Manager *
@ WTW | Thane, Maharashtra, India
Consultant Appels d'Offres Marketing Digital
@ Numberly | Paris, France