Research map / Learning theory and algorithms / Prediction and decision losses
Boosted CVaR Classification
Runtian Zhai, Chen Dan, Arun Suggala, J Zico Kolter, Pradeep Kumar Ravikumar
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use alpha-AdaLPBoost ensemble instead of single ERM model at small alpha | alpha-AdaLPBoost · COMPAS, CelebA, CIFAR-10, CIFAR-100 | test alpha-CVaR zero-one loss | Improved |
| use alpha-AdaLPBoost instead of ERM at large alpha | alpha-AdaLPBoost · all four datasets | test alpha-CVaR zero-one loss | No clear effect |
| choose model weights with LPBoost instead of uniform averaging (AdaBoost + Average) (ablation) | alpha-AdaLPBoost · CIFAR-10, CIFAR-100 | test average loss | Improved |
| choose model weights with LPBoost instead of uniform averaging (ablation) | alpha-AdaLPBoost · large alpha | alpha-CVaR zero-one loss | Improved |
| replace Regularized LPBoost (beta=100) with alpha-AdaLPBoost (ablation) | alpha-AdaLPBoost · CIFAR-10 | alpha-CVaR zero-one loss | No clear effect |
| increase number of base models T beyond 30 (ablation) | alpha-AdaLPBoost · CIFAR-10, CIFAR-100 | test alpha-CVaR zero-one loss | No clear effect |
| train ensemble with T=30 base models | AdaBoost · training cost | training time relative to ERM: 3-6 times the training time of ERM | Hurt |
Closest papers by meaning
- A Model-Based Method for Minimizing CVaR and Beyond (ICML 2023 · similarity 0.87)
- Return Capping: Sample Efficient CVaR Policy Gradient Optimisation (ICML 2025 · similarity 0.85)
- The Many Faces of Optimal Weak-to-Strong Learning (NeurIPS 2024 · similarity 0.84)
- Long-Tailed Learning as Multi-Objective Optimization (AAAI 2024 · similarity 0.83)
- Robust Minimax Boosting with Performance Guarantees (NeurIPS 2025 · similarity 0.83)
- Soft ascent-descent as a stable and flexible alternative to flooding (NeurIPS 2024 · similarity 0.83)
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.
