Research map / Learning theory and algorithms / Prediction and decision losses
Robustness of Bayesian Pool-Based Active Learning Against Prior Misspecification
Nguyen Cuong, Nan Ye, Wee Lee
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 active learning instead of passive learning with mixture prior | max Gibbs error AL · 20 Newsgroups average | AUC: 78.86 -> 82.71 | Improved |
| use active learning instead of passive learning with mixture prior | max Gibbs error AL · UCI average | AUC: 74.06 -> 76.59 | Improved |
| use bad regularizer 1/sigma^2 = 10 instead of 0.01 | max Gibbs error AL · 20 Newsgroups average | AUC: 83.37 -> 73.48 | Hurt |
| use bad regularizer 1/sigma^2 = 10 instead of 0.01 | max Gibbs error AL · UCI average | AUC: 77.05 -> 72.72 | Hurt |
| mixture prior instead of 1/sigma^2 = 1 | max Gibbs error AL · 20 Newsgroups average | AUC: 79.02 -> 82.71 | Improved |
| mixture prior instead of 1/sigma^2 = 10 | max Gibbs error AL · UCI average | AUC: 72.72 -> 76.59 | Improved |
| mixture prior instead of best single regularizer 0.01 | max Gibbs error AL · 20 Newsgroups average | AUC: 83.37 -> 82.71 | Hurt |
| mixture prior instead of 1/sigma^2 = 0.01 | max Gibbs error AL · Mushroom | AUC: 90.73 -> 90.33 | No clear effect |
| mixture prior instead of 1/sigma^2 = 0.01 | max Gibbs error AL · Ionosphere | AUC: 82.80 -> 82.95 | Improved |
| mixture prior with active learning instead of passive learning | max Gibbs error AL · Breast cancer | AUC: 89.07 -> 88.14 | Hurt |
| mixture prior with active learning instead of passive learning | max Gibbs error AL · Sonar | AUC: 66.53 -> 65.00 | Hurt |
| mixture prior with active learning instead of passive learning | max Gibbs error AL · Mushroom | AUC: 73.38 -> 90.33 | Improved |
Closest papers by meaning
- Towards Robust and Reproducible Active Learning Using Neural Networks (CVPR 2022 · similarity 0.86)
- Uncertainty-aware Active Learning for Optimal Bayesian Classifier (ICLR 2021 · similarity 0.86)
- Boosting Barely Robust Learners: A New Perspective on Adversarial Robustness (NeurIPS 2022 · similarity 0.85)
- Transductive Active Learning: Theory and Applications (NeurIPS 2024 · similarity 0.84)
- Regularized Q-learning through Robust Averaging (ICML 2024 · similarity 0.84)
- Robust Sampling for Active Statistical Inference (NeurIPS 2025 · similarity 0.84)
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.
