Research map / Probabilistic and scientific modeling / Causal discovery
Identifiability with hidden confounders
Prove and exploit conditions for discovering causal structure under latent confounding and measurement error.
85 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Causal discovery, keyword cluster “causal discovery · causal structure · graphs”
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.
21 papers in 2023–24 and 36 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 9.1% to 10.8%.
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
- 2016 · Ancestral Causal Inference (NeurIPS 2016 · 16 citations)
- 2017 · Causal Discovery Using Regression-Based Conditional Independence Tests (AAAI 2017 · 34 citations)
- 2018 · Causal Inference via Kernel Deviance Measures (NeurIPS 2018 · 30 citations)
- 2019 · Triad Constraints for Learning Causal Structure of Latent Variables (NeurIPS 2019 · 12 citations)
- 2020 · Generalized Independent Noise Condition for Estimating Latent Variable Causal Graphs (NeurIPS 2020 · 11 citations)
- 2021 · Testing Independence Between Linear Combinations for Causal Discovery (AAAI 2021 · 15 citations)
- 2022 · Root Cause Analysis of Failures in Microservices through Causal Discovery (NeurIPS 2022 · 33 citations)
- 2023 · Vector Causal Inference between Two Groups of Variables (AAAI 2023 · 5 citations)
- 2024 · Identification of Causal Structure in the Presence of Missing Data with Additive Noise Model (AAAI 2024 · 8 citations)
- 2025 · Local Causal Discovery Without Causal Sufficiency (AAAI 2025 · 2 citations)
- 2026 · Robust Causal Discovery Under Imperfect Structural Constraints (AAAI 2026 · 1 citation)
Earliest papers here
- Ancestral Causal Inference (NeurIPS 2016 · 16 citations)
- Causal Discovery Using Regression-Based Conditional Independence Tests (AAAI 2017 · 34 citations)
- Causal Inference via Kernel Deviance Measures (NeurIPS 2018 · 30 citations)
What has been tried
61 papers here report 451 measured changes (a component added, removed or swapped, compared with a baseline): helped 70%, no clear effect 10%, hurt 15%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| vary the assumed delay distribution from exponential to other distributions | Synthetic data, distribution misspecification · F1 | 0.82; 0.81; 0.76; 0.75 | Mixed | Causal Discovery from Event Sequences by Local Cause-Effect Attribution (NeurIPS 2024 · 1 citation) |
| Omit MINE-based independence enforcement for exogenous variables | Image data · final output | No clear effect | Differentiable Causal Discovery for Latent Hierarchical Causal Models (ICLR 2025 · 0 citations) | |
| Add symmetric CI post-processing to symmetric CI discovery | Causal discovery, d=3 · SHD (× 10^2) | 57 ± 84 -> 13 ± 4 | Helped | Signature Kernel Conditional Independence Tests in Causal Discovery for Stochastic Processes (ICLR 2025 · 0 citations) |
| Use IDI's in-distribution interventions rather than counterfactuals in step 2 | Synthetic RQ2 Non-Lin Invertible, multiple root causes · Recall@1 | 0.77 -> 0.83 | Helped | Robust Root Cause Diagnosis using In-Distribution Interventions (ICLR 2025 · 0 citations) |
| Replace fixed median-heuristic kernel bandwidth with trainable kernel parameters while retaining conditional likelihood | Toy example · Causal discovery accuracy | Hurt | Optimal Kernel Choice for Score Function-based Causal Discovery (ICML 2024 · 0 citations) | |
| Increase bin number and sample size | Synthetic data, confounding and mediation settings · type I error rate | Helped | Causal Discovery via Conditional Independence Testing with Proxy Variables (ICML 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 27 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- Sachs flow cytometry dataset · 2 of 27 papers
Baselines
- KCI · 3 of 27 papers
- PC · 3 of 27 papers
- CAM · 3 of 27 papers
- FCI · 2 of 27 papers
- NOTEARS · 2 of 27 papers
- NOGAM · 2 of 27 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 57 papers; from OpenAlex and the papers' affiliations).
- Mohamed bin Zayed University of Artificial Intelligence · 12 papers
- Carnegie Mellon University · 12 papers
- Shantou University · 5 papers
- Guangdong University of Technology · 5 papers
- University of California, San Diego · 4 papers
- CISPA Helmholtz Center for Information Security · 4 papers
- Peng Cheng Laboratory · 3 papers
- University of Melbourne · 3 papers
- Fudan University · 3 papers
- University of Oxford · 3 papers
Use it
Download BibTeX of 38 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
- Constraint-based structure learning with latents (90 papers)
- Interventional causal discovery (68 papers)
- Differentiable and score-based DAG learning (52 papers)
- Causal discovery from time series (51 papers)
Elsewhere in the topic
- Causal representation identifiability (122 papers)
- Latent-variable causal estimation (117 papers)
- Causal effect identification (107 papers)
- Counterfactual identification and bounds (92 papers)
- Causal approaches to fairness (46 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).