Research map / Label-efficient and robust learning / Test-time adaptation
Robust optimization for distribution shift
Trains for robustness via risk extrapolation, selective augmentation and adaptive risk minimization.
34 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2020–2026) · an approach in Test-time adaptation, keyword cluster “shift · distribution · machine”
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.
10 papers in 2023–24 and 9 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 9.1% to 5.1%.
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 · Improving robustness against common corruptions by covariate shift adaptation (NeurIPS 2020 · 158 citations)
- 2021 · Out-of-Distribution Generalization via Risk Extrapolation (REx) (ICML 2021 · 251 citations)
- 2022 · Wild-Time: A Benchmark of in-the-Wild Distribution Shift over Time (NeurIPS 2022 · 22 citations)
- 2023 · Can You Rely on Your Model Evaluation? Improving Model Evaluation with Synthetic Test Data (NeurIPS 2023 · 5 citations)
- 2024 · Reconstruct and Match: Out-of-Distribution Robustness via Topological Homogeneity (NeurIPS 2024 · 2 citations)
- 2025 · Sufficient Invariant Learning for Distribution Shift (CVPR 2025 · 3 citations)
- 2026 · Mitigating Spurious Correlation via Distributionally Robust Learning with Hierarchical Ambiguity Sets (ICLR 2026 · 0 citations)
Earliest papers here
- Improving robustness against common corruptions by covariate shift adaptation (NeurIPS 2020 · 158 citations)
- Robustness to Spurious Correlations via Human Annotations (ICML 2020 · 25 citations)
- Out-of-Distribution Generalization via Risk Extrapolation (REx) (ICML 2021 · 251 citations)
What has been tried
29 papers here report 284 measured changes (a component added, removed or swapped, compared with a baseline): helped 58%, no clear effect 12%, hurt 21%, mixed 8%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add SSR to REMA | VLCS · accuracy (%) | 78.3 -> 79.4 | Helped | Reconstruct and Match: Out-of-Distribution Robustness via Topological Homogeneity (NeurIPS 2024 · 2 citations) |
| use the first quantile statistic Φq instead of the second statistic Φ²q | Folktables · FDP | 0.019 -> 0.004 | Helped | Sequential Harmful Shift Detection Without Labels (NeurIPS 2024 · 1 citation) |
| Remove the second term from the proposed loss | MNIST, varying πtr · test AUC | Hurt | AUC Maximization under Positive Distribution Shift (NeurIPS 2024 · 0 citations) | |
| Add Gumbel-softmax routing instead of direct softmax | Kinetics50-C video shift: Gaussian noise · accuracy | 52.0 -> 52.6 | Helped | Test-Time Selective Adaptation for Uni-Modal Distribution Shift in Multi-Modal Data (ICML 2025) |
What reviewers will expect
Datasets and baselines named in the 7 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- OfficeHome · 2 of 7 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 19 papers; from OpenAlex and the papers' affiliations).
- Google · 2 papers
- Meta · 2 papers
- Nanjing University · 2 papers
Use it
Download BibTeX of 31 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
- Characterizing out-of-distribution failures (28 papers)
Elsewhere in the topic
- Pseudo-label test-time adaptation (90 papers)
- Test-time adaptation of vision-language models (79 papers)
- Realistic and stable test-time adaptation (71 papers)
- Extrapolation and single-sample adaptation (57 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).