Research map / Learning theory and algorithms / Prediction and decision losses
PAC Confidence Sets for Deep Neural Networks via Calibrated 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 |
|---|---|---|---|
| Add temperature scaling to the probability forecaster (ablation) | ResNet PAC confidence sets · ImageNet | maximum confidence set size | Helped |
| Use the direct generalization bound instead of the VC bound (ablation) | PAC confidence sets · ImageNet | confidence set size | Helped |
| Add temperature scaling to the probability forecaster (ablation) | visual tracking PAC confidence sets · Visual object tracking benchmark | confidence set size | Helped |
| Use the direct generalization bound instead of the VC bound (ablation) | visual tracking PAC confidence sets · Visual object tracking benchmark | confidence set size | Helped |
| Add temperature scaling to the probability forecaster (ablation) | half-cheetah trajectory confidence sets · Half-cheetah | confidence set size | Helped |
| Accumulate predicted variances over time (ablation) | half-cheetah trajectory confidence sets · Half-cheetah | confidence set size | Helped |
| Use the direct generalization bound instead of the VC bound (ablation) | half-cheetah trajectory confidence sets · Half-cheetah | confidence set size | Helped |
Closest papers by meaning
- PAC Confidence Predictions for Deep Neural Network Classifiers (ICLR 2021 · similarity 0.90)
- Certifying Deep Network Risks and Individual Predictions with PAC-Bayes Loss via Localized Priors (NeurIPS 2025 · similarity 0.85)
- Pointwise Information Measures as Confidence Estimators in Deep Neural Networks: A Comparative Study (ICML 2025 · similarity 0.83)
- SURE: SUrvey REcipes for building reliable and robust deep networks (CVPR 2024 · similarity 0.83)
- A Benchmark Study on Calibration (ICLR 2024 · similarity 0.82)
- PAC Prediction Sets Under Label Shift (ICLR 2024 · similarity 0.82)
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.