Research map / Learning theory and algorithms / Sample complexity
Constants Matter: The Performance Gains of Active Learning
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 |
|---|---|---|---|
| Adaptive uncertainty-sampling variant (half random, half uncertainty gradient steps) instead of random MLE sampling | well-specified logistic regression on sphere · theory, expected excess error | excess error rate in n | No clear effect |
| Adaptive sampling (Algorithm 1) instead of random sampling | logistic regression on sphere · theory, excess error bound | upper bound constant dependence on err*: adaptive bound scales with err*·d/n vs random d log d/n | Helped |
| Adaptive sampling vs random sampling on excess error scaling with Bayes error (ablation) | logistic regression on sphere, synthetic experiments · Figure 4 (M varied, n=1000, d=10) | test excess error vs limiting error | Mixed |
| Random sampling MLE in high dimension (ablation) | logistic regression on sphere · Figure 5 (n=100·d, r=1, M=20) | test excess error vs dimension | No clear effect |
Closest papers by meaning
- Transductive Active Learning: Theory and Applications (NeurIPS 2024 · similarity 0.87)
- Uncertainty for Active Learning on Graphs (ICML 2024 · similarity 0.86)
- Robust Sampling for Active Statistical Inference (NeurIPS 2025 · similarity 0.86)
- On the Benefits of Active Data Collection in Operator Learning (ICML 2025 · similarity 0.86)
- Uncertainty Herding: One Active Learning Method for All Label Budgets (ICLR 2025 · similarity 0.86)
- On the Convergence of Loss and Uncertainty-based Active Learning Algorithms (NeurIPS 2024 · similarity 0.85)
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.