Research map / Label-efficient and robust learning / Self-supervised and contrastive
GenView: Enhancing View Quality with Pretrained Generative Model for Self-Supervised Learning
Xiaojie Li, Yibo Yang, Xiangtai Li, Jianlong Wu, Yue Yu, Bernard Ghanem, Min Zhang
This paper tried a common change: Add a contrastive loss
add quality-driven contrastive loss to framework · IN-100 linear eval · top-1 accuracy: 71.50 -> 74.88 Improved
What 1,317 accepted papers report for this change
- 1,253 improved
- 22 no clear effect
- 16 mixed
- 26 hurt
- Beyond Looking Up, Try Looking Around: Harmonizing Global Structure and Local Consistency in Optimal Transport for Short Text Clustering ICML 2026 · add contrastive learning · SearchSnippets · ACC: 56.65 -> 80.12Improved
- SleepLM: Natural-Language Intelligence for Human Sleep ICML 2026 · add contrastive objective to captioning (CoCa vs Cap) · sleep event IoU · IoU: 29.2 -> 27.4Hurt
- LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations ICML 2026 · add contrastive loss L_con · FS1000 average · Spearman correlation: 0.759 -> 0.789Improved
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 GenView to MoCov2 | MoCov2 ResNet-50 · ImageNet-1K linear eval | top-1 accuracy: 67.5 -> 70.0 | Improved |
| add GenView to MoCov2 | MoCov2 ResNet-50 · ImageNet-1K semi-supervised 1% labels | top-1 accuracy: 42.1 -> 50.6 | Improved |
| add GenView to BYOL | BYOL ResNet-50 · MS-COCO detection | AP: 40.6 -> 41.2 | Improved |
| add 0.3M Laion400M images to ImageNet-1K pretraining (ablation) | MoCov3 ResNet-50 · ImageNet-1K linear eval (50 ep) | top-1 accuracy: 62.39 -> 63.31 | Improved |
| use GenView framework with 0.15M generated images instead of adding 0.3M extra real or synthetic images (ablation) | MoCov3 ResNet-50 · ImageNet-1K linear eval (50 ep) | top-1 accuracy: 62.39 -> 65.62 | Improved |
| add generated-view framework without adaptive noise or quality loss (ablation) | MoCov3 ResNet-18 · IN-100 linear eval | top-1 accuracy: 65.52 -> 71.50 | Improved |
| add adaptive view generation to framework (ablation) | GenView · IN-100 linear eval | top-1 accuracy: 71.50 -> 73.96 | Improved |
| add quality-driven contrastive loss to framework (ablation) | GenView · IN-100 linear eval | top-1 accuracy: 71.50 -> 74.88 | Improved |
| add quality-driven contrastive loss to baseline without generated views (ablation) | MoCov3 ResNet-18 · IN-100 linear eval | top-1 accuracy: 65.52 -> 66.97 | Improved |
| use adaptive noise level selection instead of constant or random (ablation) | GenView · IN-100 linear eval | top-1 accuracy: 72.96 -> 73.96 | Improved |
| raise noise level constant from 0 to 100-400 (ablation) | GenView · IN-100 linear eval | top-1 accuracy: 71.80 -> 72.14 (CS 100) | No clear effect |
| increase probability of applying GenView (ablation) | GenView · ImageNet-1K linear eval (50 ep) | top-1 accuracy: 62.39 -> 70.55 | Improved |
Closest papers by meaning
- Can Generative Models Improve Self-Supervised Representation Learning? (AAAI 2025 · similarity 0.88)
- GenAI Arena: An Open Evaluation Platform for Generative Models (NeurIPS 2024 · similarity 0.86)
- Beyond Random Augmentations: Pretraining with Hard Views (ICLR 2025 · similarity 0.86)
- Understand Before You Generate: Self-Guided Training for Autoregressive Image Generation (NeurIPS 2025 · similarity 0.86)
- Personalized Representation from Personalized Generation (ICLR 2025 · similarity 0.86)
- GenWarp: Single Image to Novel Views with Semantic-Preserving Generative Warping (NeurIPS 2024 · similarity 0.86)
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.
