machine-learning-privacy

A curated list of Federated Learning papers/articles and recent advancements.

Official implementation of "Graph Unlearning" (ACM CCS 2022)

A School for All Seasons on Trustworthy Machine Learning

Official implementation of "FACE-AUDITOR: Data Auditing in Facial Recognition Systems" (USENIX Security 2023)

The source code of the paper "Learning-Based Difficulty Calibration for Enhanced Membership Inference Attacks", in IEEE European Symposium on Security and Privacy (EuroS&P 2024)

There are many popular fitness tracking apps on the market, usually joined with the purchase of their respective hardware (wearable band, smartwatch). The dark side of these apps and wearable devices is about information privacy, as society considers medical and biometric data highly sensitive.

Code for our paper 'Formalizing Distribution Inference Risks'

Official AISTATS 2026 implementation formalizing Sensitivity Awareness (SA) via Differential Privacy. Features a 4-bit LoRA post-training pipeline to secure LLM agents against contextual authorization and access-control bypasses.