Research map / Label-efficient and robust learning / Recognition and long-tail
ACLS: Adaptive and Conditional Label Smoothing for Network Calibration
Hyekang Park, Jongyoun Noh, Youngmin Oh, Donghyeon Baek, Bumsub Ham
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 ACLS adaptive smoothing function only instead of MDCA adaptive regularization (ablation) | ACLS · Tiny-ImageNet, ResNet-50 | ECE (%): 2.77 -> 1.54 | Improved |
| use ACLS indicator function only instead of MbLS (ablation) | ACLS · Tiny-ImageNet, ResNet-50 | ECE (%): 1.64 -> 1.32 | Improved |
| combine adaptive smoothing and conditional indicator (ablation) | ACLS · Tiny-ImageNet, ResNet-50 | ECE (%): 1.32 -> 1.05 | Improved |
| use margin condition instead of ranking condition (ablation) | ACLS · Tiny-ImageNet, ResNet-50 | ECE (%): 1.48 -> 1.05 | Improved |
| use margin condition instead of ranking condition (ablation) | CRL · Tiny-ImageNet, ResNet-50 | ECE (%): 1.65 -> 1.47 | Improved |
| ACLS loss versus CE | ResNet-50 · ImageNet val | ECE (%): 9.10 -> 1.02 | Improved |
| ACLS versus MbLS on segmentation accuracy | DeepLabV3 ResNet-50 · PASCAL VOC val | mIoU: 71.20 -> 70.63 | Hurt |
| ACLS versus label smoothing on Tiny-ImageNet accuracy | ResNet-50 · Tiny-ImageNet val | top-1 accuracy: 65.78 -> 64.84 | Hurt |
| ACLS regulation of top logit can flip predictions | ACLS · Tiny-ImageNet | samples with altered prediction: 0.29% of samples | Hurt |
Later papers that compared with this method
1 later accepted paper reports ACLS as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- FedLAS: Feature-Modulated Bidirectional Label Smoothing for Neural Network Calibration (ECCV 2026 · better on 3 of 3 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MaxSup: Overcoming Representation Collapse in Label Smoothing (NeurIPS 2025 · similarity 0.87)
- Towards Calibrated Multi-label Deep Neural Networks (CVPR 2024 · similarity 0.86)
- Training Deep Neural Networks with Virtual Smoothing Classes (AAAI 2025 · similarity 0.86)
- Improving GNN Calibration with Discriminative Ability: Insights and Strategies (AAAI 2024 · similarity 0.86)
- Moderate Message Passing Improves Calibration: A Universal Way to Mitigate Confidence Bias in Graph Neural Networks (AAAI 2024 · similarity 0.85)
- A Benchmark Study on Calibration (ICLR 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.
