2 papers
cs.CV2026
Spectral Gradient Orthogonalization Improves Differentially Private Training at Scale
Sabari Shanmugam, Nick Barnes, Kerry Taylor
Differentially private training adds isotropic Gaussian noise to clipped gradients, corrupting every singular direction equally. In vision models, where spatial correlation concent…
cs.LG2022
Implicit Mixture of Interpretable Experts for Global and Local Interpretability
Nathan Elazar, Kerry Taylor
We investigate the feasibility of using mixtures of interpretable experts (MoIE) to build interpretable image classifiers on MNIST10. MoIE uses a black-box router to assign each in…