Research map / Label-efficient and robust learning / Self-supervised and contrastive
Empirical contrastive visual learning
Frameworks and studies of when contrastive visual and dense representation learning works.
38 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2020–2025) · an approach in Self-supervised and contrastive, keyword cluster “supervised contrastive · contrastive self · negative”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
12 papers in 2023–24 and 3 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.7% to 0.9%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2020 · Momentum Contrast for Unsupervised Visual Representation Learning (CVPR 2020 · 12,585 citations)
- 2021 · Exploring Simple Siamese Representation Learning (CVPR 2021 · 3,493 citations)
- 2022 · When Does Contrastive Visual Representation Learning Work? (CVPR 2022 · 99 citations)
- 2023 · MaskCon: Masked Contrastive Learning for Coarse-Labelled Dataset (CVPR 2023 · 17 citations)
- 2024 · ODCR: Orthogonal Decoupling Contrastive Regularization for Unpaired Image Dehazing (CVPR 2024 · 31 citations)
- 2025 · Rethinking Masked Data Reconstruction Pretraining for Strong 3D Action Representation Learning (AAAI 2025 · 2 citations)
Earliest papers here
- Momentum Contrast for Unsupervised Visual Representation Learning (CVPR 2020 · 12,585 citations)
- A Simple Framework for Contrastive Learning of Visual Representations (ICML 2020 · 7,181 citations)
- Unsupervised Learning of Visual Features by Contrasting Cluster Assignments (NeurIPS 2020 · 2,015 citations)
What has been tried
35 papers here report 422 measured changes (a component added, removed or swapped, compared with a baseline): helped 69%, no clear effect 7%, hurt 22%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add orthogonal decoupling optimized on the Stiefel manifold | SOTS-indoor · PSNR (dB) | 22.30 -> 26.32 | Helped | ODCR: Orthogonal Decoupling Contrastive Regularization for Unpaired Image Dehazing (CVPR 2024 · 31 citations) |
| Add Variance/Covariance terms | ImageNet-1K linear probing · linprob | 63.7 -> 68.3 | Helped | Connecting Joint-Embedding Predictive Architecture with Contrastive Self-supervised Learning (NeurIPS 2024 · 8 citations) |
| Add WCD while SCR and DCAS are omitted | nuScenes-lidarseg 1% · mIoU | 37.89 -> 42.33 | Helped | Fine-grained Image-to-LiDAR Contrastive Distillation with Visual Foundation Models (NeurIPS 2024 · 3 citations) |
| Apply NAT to the masked joint reconstruction target instead of using the raw joint target | NTU-60 X-sub · top-1 accuracy | 74.8 -> 85.1 | Helped | Rethinking Masked Data Reconstruction Pretraining for Strong 3D Action Representation Learning (AAAI 2025 · 2 citations) |
| use similar-dimension clusters as masking units instead of randomly aggregating dimensions | Permuted CIFAR-10 · probing accuracy | 55.7% -> 86.4% | Helped | URLOST: Unsupervised Representation Learning without Stationarity or Topology (ICLR 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- MAE · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 15 papers; from OpenAlex and the papers' affiliations).
- University of California, Berkeley · 2 papers
- Microsoft · 2 papers
- New York University · 2 papers
Use it
Download BibTeX of 25 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Contrastive versus non-contrastive theory (74 papers)
- Inductive bias and minimal representations (66 papers)
- Representation evaluation and trade-offs (37 papers)
Elsewhere in the topic
- Invariance-based self-supervised learning (168 papers)
- Pretext-task visual pre-training (136 papers)
- Data-efficient and robust self-supervision (122 papers)
- Self-supervision in supervised pipelines (121 papers)
- Self-supervised neuroimaging representations (98 papers)
- Self-supervised speech and brain decoding (56 papers)
- Masked autoencoder pretraining (48 papers)
- Self-supervised speech representations (24 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).