Research map / Optimization / Nonconvex and smooth optimization
Bregman Divergence for Stochastic Variance Reduction: Saddle-Point and Adversarial Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Replace squared Euclidean Bregman divergence with KL (entropic) divergence in Breg-SVRG (ablation) | adversarial prediction (F-score) with n-by-n variables · synthetic and ionosphere datasets | primal gap vs number of effective passes | Helped |
| Replace squared Euclidean Bregman divergence with KL (entropic) divergence in Breg-SVRG (ablation) | entropy regularized LPBoost · adult | primal gap vs number of effective passes | Helped |
| Use KL divergence instead of Euclidean in Breg-SVRG (ablation) | adversarial prediction F-score · synthetic and ionosphere | test F-score vs CPU time | Helped |
| Use KL divergence instead of Euclidean in Breg-SVRG (ablation) | LPBoost test accuracy · adult | test accuracy vs #pass | Helped |
| Add universal catalyst with non-uniform sampling to Euclidean-SVRG (ablation) | Euclidean-SVRG on LPBoost · adult | primal gap | Helped |
| Add universal catalyst with non-uniform sampling to Entropy-SVRG (ablation) | Entropy-SVRG on LPBoost · adult | primal gap | No clear effect |
| Replace KL (entropic) divergence with Euclidean in the scheme (a) factorization (per-entry bilinear factor) | adversarial prediction smoothness · theoretical comparison | smoothness constant L1 vs L2 | No clear effect |
Closest papers by meaning
- Learning Bregman Divergences with Application to Robustness (NeurIPS 2024 · similarity 0.89)
- Stabilizing Adversarial Nets with Prediction Methods (ICLR 2018 · similarity 0.87)
- Between Stochastic and Adversarial Online Convex Optimization: Improved Regret Bounds via Smoothness (NeurIPS 2022 · similarity 0.87)
- Adversarial Multiclass Classification: A Risk Minimization Perspective (NeurIPS 2016 · similarity 0.87)
- Parameter-free Algorithms for the Stochastically Extended Adversarial Model (NeurIPS 2025 · similarity 0.86)
- Kernel Learning with Adversarial Features: Numerical Efficiency and Adaptive Regularization (NeurIPS 2025 · similarity 0.86)
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.