Badih Ghazi

Fixed-Parameter Tractability of Private Synthetic Data Generation COLT 2026 Nearly Linear-Time User-Level DP-SCO with Optimal Rates COLT 2026 Balls-and-Bins Sampling for DP-SGD AISTATS 2025 On Memorization of Large Language Models in Logical Reasoning IJCNLP 2025 PREM: Privately Answering Statistical Queries with Relative Error COLT 2025 Scaling Laws for Differentially Private Language Models ICML 2025 Unlearn and Burn: Adversarial Machine Unlearning Requests Destroy Model Accuracy ICLR 2025 Differentially Private Optimization with Sparse Gradients NeurIPS 2024 How Private are DP-SGD Implementations? ICML 2024 Individualized Privacy Accounting via Subsampling with Applications in Combinatorial Optimization ICML 2024 LabelDP-Pro: Learning with Label Differential Privacy via Projections ICLR 2024 On Convex Optimization with Semi-Sensitive Features COLT 2024 Scalable DP-SGD: Shuffling vs. Poisson Subsampling NeurIPS 2024 Differentially Private Heatmaps AAAI 2023 On Computing Pairwise Statistics with Local Differential Privacy NeurIPS 2023 On Differentially Private Sampling from Gaussian and Product Distributions NeurIPS 2023 On User-Level Private Convex Optimization ICML 2023 Optimal Unbiased Randomizers for Regression with Label Differential Privacy NeurIPS 2023 Regression with Label Differential Privacy ICLR 2023 Sparsity-Preserving Differentially Private Training of Large Embedding Models NeurIPS 2023 Ticketed Learning–Unlearning Schemes COLT 2023 User-Level Differential Privacy With Few Examples Per User NeurIPS 2023 Anonymized Histograms in Intermediate Privacy Models NeurIPS 2022 Faster Privacy Accounting via Evolving Discretization ICML 2022 Large-Scale Differentially Private BERT EMNLP 2022 Private Isotonic Regression NeurIPS 2022 Private Rank Aggregation in Central and Local Models AAAI 2022 Deep Learning with Label Differential Privacy NeurIPS 2021 Differentially Private Aggregation in the Shuffle Model: Almost Central Accuracy in Almost a Single Message ICML 2021 Locally Private k-Means in One Round ICML 2021 Near-tight Closure Bounds for the Littlestone and Threshold Dimensions ALT 2021 On Avoiding the Union Bound When Answering Multiple Differentially Private Queries COLT 2021 Robust and Private Learning of Halfspaces AISTATS 2021 User-Level Differentially Private Learning via Correlated Sampling NeurIPS 2021 Differentially Private Clustering: Tight Approximation Ratios NeurIPS 2020 Private Counting from Anonymous Messages: Near-Optimal Accuracy with Vanishing Communication Overhead ICML 2020 Recursive Sketches for Modular Deep Learning ICML 2019