🏢 MIT-IBM Watson AI Lab, IBM Research
Privacy without Noisy Gradients: Slicing Mechanism for Generative Model Training
·1432 words·7 mins·
loading
·
loading
AI Theory
Privacy
🏢 MIT-IBM Watson AI Lab, IBM Research
Train high-quality generative models with strong differential privacy using a novel slicing mechanism that injects noise into random low-dimensional data projections, avoiding noisy gradients.