Statistical Learning
Robustness of Bayesian Neural Networks to Gradient-Based Attacks
Despite significant efforts, both practical and theoretical, the problem remains open.
90fd4f88f588ae64038134f1eeaa023f-Paper.pdf
Previous coreset constructions only allow one missing coordinate.
PrivatelyPublishablePer-instancePrivacy
We consider how to privately share the personalized privacy losses incurred by objective perturbation, using per-instance differential privacy (pDP).
Evaluatingmodelperformanceunderworst-case subpopulations
The training population typically does not accurately represent what the model will encounter underoperation.
