Explore a novel approach to differentially private deep learning in this 17-minute IEEE conference talk. Delve into the concept of Spectral-DP, which utilizes spectral perturbation and filtering techniques to enhance privacy in machine learning models. Learn from researchers at Lehigh University and the University of Connecticut as they present their innovative method for balancing model accuracy and data protection in deep learning applications.
Spectral-DP - Differentially Private Deep Learning through Spectral Perturbation and Filtering
Overview
Syllabus
Spectral-DP: Differentially Private Deep Learning through Spectral Perturbation and Filtering
Taught by
IEEE Symposium on Security and Privacy
Tags
Related Courses
-
Locally Differentially Private Sparse Vector Aggregation
-
Learning Differentially Private Mechanisms
-
Locally Differentially Private Frequency Estimation Based on Convolution Framework
-
Towards Practical Differentially Private Convex Optimization
-
Is Interaction Necessary for Distributed Private Learning?
-
A Theory to Instruct Differentially-Private Learning via Clipping Bias Reduction
