AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Scaling Up Differentially Private LASSO Regularized Logistic Regression via Faster Frank-Wolfe Iterations

Edward Raff, Amol Ashish Khanna, Fred Lu

NeurIPS 2023 · Federated learning and privacy › Differential privacy · approach: Private online and user-level learning · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use the Big-Step Little-Step sampler instead of brute-force noisy-max (ablation)Sparse DP Frank-Wolfe · News20, ε=1speedup: 17.83 -> 81.69Improved
Use the Big-Step Little-Step sampler instead of brute-force noisy-max (ablation)Sparse DP Frank-Wolfe · URL, ε=1speedup: 1.02 -> 9.99Improved
Use the Big-Step Little-Step sampler instead of brute-force noisy-max (ablation)Sparse DP Frank-Wolfe · RCV1, ε=0.1speedup: 1.82 -> 20.37Improved
Use the Big-Step Little-Step sampler instead of brute-force noisy-max (ablation)Sparse DP Frank-Wolfe · Web, ε=1speedup: 21.24 -> 581.25Improved
Use the Big-Step Little-Step sampler instead of brute-force noisy-max (ablation)Sparse DP Frank-Wolfe · KDDA, ε=0.1speedup: 368.96 -> 2245.56Improved
Apply sparse-aware mathematically equivalent Frank-Wolfe updatesNon-private Frank-Wolfe · RCV1, News20, URL, Web, and KDDAtest accuracyNo clear effect
Use the sparse non-private Frank-Wolfe implementationNon-private Frank-Wolfe · Sparse datasetsruntimeNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy