Research map / Architectures and efficiency / RNNs and normalization
Boost Neural Networks by Checkpoints
Feng Wang, Guoyizhe Wei, Qiao Liu, Jinxiang Ou, Xian Wei, Hairong Lv
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 adaptive sample reweighting to train and ensemble checkpoints instead of the strongest checkpoint-ensemble baseline | ResNet-110 · Cifar-100 | error rate (%): 24.27 ± 0.26 -> 23.51 ± 0.20 | Improved |
| Use CBNN rather than the strongest checkpoint-ensemble baseline | DenseNet-40 · Tiny-ImageNet | error rate (%): 36.35 ± 0.30 -> 34.28 ± 0.20 | Improved |
| Use CBNN rather than the strongest checkpoint-ensemble baseline | EfficientNet-B0 · Cifar-100 | error rate (%): 19.58 ± 0.24 -> 18.06 ± 0.20 | Improved |
| Use CBNN rather than the strongest checkpoint-ensemble baseline | EfficientNet-B3 · Cifar-10 | error rate (%): 3.27 -> 3.28 | Mixed |
| Use CBNN rather than a single model | EfficientNet-B0 · Cifar-100 (imbalanced) | error rate (%): 29.19 -> 24.39 | Improved |
| Use CBNN rather than a single model | EfficientNet-B3 · Tiny-ImageNet (imbalanced) | error rate (%): 43.34 -> 39.02 | Improved |
| Use CBNN instead of a single model to reach the 70% accuracy target | EfficientNet-B0 · Cifar-100 | GPU hours: 7.47 ± 0.79 -> 4.82 ± 0.22 | Improved |
Closest papers by meaning
- Self-Ensemble Protection: Training Checkpoints Are Good Data Protectors (ICLR 2023 · similarity 0.88)
- Ranking Neural Checkpoints (CVPR 2021 · similarity 0.86)
- Optimal Gradient Checkpoint Search for Arbitrary Computation Graphs (CVPR 2021 · similarity 0.86)
- United We Stand: Using Epoch-Wise Agreement of Ensembles to Combat Overfit (AAAI 2024 · similarity 0.84)
- When, Where and Why to Average Weights? (ICML 2025 · similarity 0.83)
- Repeated Random Sampling for Minimizing the Time-to-Accuracy of Learning (ICLR 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.
