Iman Sharei·1d agoDifferential Privacy for Quantum ClassifiersWhat DP-SGD costs a variational circuit, from a single model to a hundred federated clients
Adnan Masood, PhD.·6d agoPrivacy-Preserving AI Text Detection: Estimating LLM Usage with Differential PrivacyHow Stanford’s Red-Green Sequence watermark limits the evidence in one text while preserving a useful signal across many.
InData Science CollectivebyMinha Hwang·Nov 1Improving Enterprise AI and Agent When You Can’t See the DataFour ways to turn eyes-off enterprise data into insights, model updates, evaluations and post-training data, and when each fits chat…
Sam Austin AI·Oct 9Differential Privacy Tutorial: Train Private Models with OpacusA model that scores well on your test set can still leak details about its training data, and anyone who can query it can probe for them…
Ezio Chifunga·Oct 6Protecting the Person Without Erasing InequalityWhat we learned building a privacy and fairness middleware for Brazil public health data and why anonymization can make minorities disappear
Howard Wong·Aug 8Protecting the People Behind the NumbersStatistical Disclosure Control and Data Privacy in National Statistical OfficesA response icon1A response icon1
Howard Wong·Aug 7How Differential Privacy WorksA technical appendix on the implementation of differential privacy in official statistics.
Swethasrihari·Jul 23Beyond Anonymization: Understanding Differential PrivacyPart 3 of the Data Privacy Series
