Research map / Label-efficient and robust learning / Semi-supervised learning
Webly Supervised Image Classification with Self-Contained Confidence
Jingkang Yang, Litong Feng, Weirong Chen, Xiaopeng Yan, Huabin Zheng, Ping Luo, Wayne Zhang
This paper tried a common change: Add dropout
add MC dropout to pretrained model · WebVision-500 (stage 2) · Top-1 accuracy (%): 76.46 -> 76.00 No clear effect
What 795 accepted papers report for this change
- 549 improved
- 87 no clear effect
- 12 mixed
- 147 hurt
- Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference ICML 2026 · apply layer dropout to whole transformer blocks instead of attention and FFN sub-blocks separately · 20 TPP val loss, 271M · val loss: 2.377 -> 2.367Improved
- Quantifying the Uncertainty of Foundation Models with Singular Value Ensembles ICML 2026 · use MC Dropout · vision and SST-2 · accuracy: 86.3 -> 76.6 on Flowers102Hurt
- Motion Planning in Compressed Representation Spaces ICML 2026 · add nested dropout during training · WOMD guided generation · goal FDE: 4.145 -> 1.104Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 |
|---|---|---|---|
| add graph-based aggregation (GBA) to vanilla model (ablation) | SCC framework · ImageNet val (stage 2) | Top-1 accuracy (%): 69.78 -> 71.56 | Improved |
| add GBA to vanilla model (ablation) | SCC framework · WebVision-500 (stage 2) | Top-1 accuracy (%): 76.46 -> 77.12 | Improved |
| add GBA to mixup model (ablation) | SCC framework · WebVision-500 (stage 2) | Top-1 accuracy (%): 77.47 -> 77.76 | Improved |
| add mixup regularization to pretrained model (ablation) | SCC framework · ImageNet val (stage 2) | Top-1 accuracy (%): 69.78 -> 72.25 | Improved |
| add entropy regularizer to pretrained model (ablation) | SCC framework · WebVision-500 (stage 2) | Top-1 accuracy (%): 76.46 -> 73.78 | Hurt |
| add label smoothing to pretrained model (ablation) | SCC framework · WebVision-500 (stage 2) | Top-1 accuracy (%): 76.46 -> 77.02 | Improved |
| add MC dropout to pretrained model (ablation) | SCC framework · WebVision-500 (stage 2) | Top-1 accuracy (%): 76.46 -> 76.00 | No clear effect |
| replace vanilla with ensemble (E=5) confidence provider (ablation) | SCC framework · WebVision-500 (stage 2) | Top-1 accuracy (%): 76.46 -> 79.12 | Improved |
| replace sample-wise SCC with constant confidence (ablation) | SCC framework · WebVision-500 | Top-1 accuracy (%): 77.17 (Vanilla+GBA) not exceeded by any constant | Hurt |
| use mixup as confidence provider instead of vanilla (ablation) | SCC framework · WebVision-500 SCC quality | ECE: 0.2371 -> 0.1828 | Improved |
| add GBA to vanilla confidence provider (ablation) | SCC framework · WebVision-500 SCC quality | OCE: 0.1518 -> 0.0002 | Improved |
| add GBA to vanilla confidence provider (ablation) | SCC framework · WebVision-500 SCC quality | ECE: 0.2371 -> 0.2494 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MentorNet — MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels (ICML 2018) · better on 1 of 1 shared result
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
- Web-Supervised Network with Softly Update-Drop Training for Fine-Grained Visual Classification (AAAI 2020 · similarity 0.85)
- Webly Supervised Learning Meets Zero-Shot Learning: A Hybrid Approach for Fine-Grained Classification (CVPR 2018 · similarity 0.85)
- Webly Supervised Fine-Grained Recognition: Benchmark Datasets and an Approach (ICCV 2021 · similarity 0.85)
- Attend in Groups: A Weakly-Supervised Deep Learning Framework for Learning From Web Data (CVPR 2017 · similarity 0.85)
- Learning From Web Data With Self-Organizing Memory Module (CVPR 2020 · similarity 0.85)
- Barely-Supervised Learning: Semi-supervised Learning with Very Few Labeled Images (AAAI 2022 · 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.
