Research map / Probabilistic and scientific modeling / Molecules and proteins
Counterfactual and statistical explanations
Generates counterfactual and logic-based explanations and frames explainability as inference.
81 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Molecules and proteins, keyword cluster “decision · machine · statistical”
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.
31 papers in 2023–24 and 24 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 6.3% to 2.6%.
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
- 2017 · Collaborative Company Profiling: Insights from an Employee's Perspective (AAAI 2017 · 33 citations)
- 2018 · A Robust Approach to Sequential Information Theoretic Planning (ICML 2018 · 7 citations)
- 2019 · EDDI: Efficient Dynamic Discovery of High-Value Information with Partial VAE (ICML 2019 · 43 citations)
- 2020 · Temporal Logics Over Finite Traces with Uncertainty (AAAI 2020 · 22 citations)
- 2021 · Explaining A Black-box By Using A Deep Variational Information Bottleneck Approach (AAAI 2021 · 66 citations)
- 2022 · Amortized Generation of Sequential Algorithmic Recourses for Black-Box Models (AAAI 2022 · 20 citations)
- 2023 · Blending Advertising with Organic Content in E-commerce via Virtual Bids (AAAI 2023 · 10 citations)
- 2024 · Symbolic Cognitive Diagnosis via Hybrid Optimization for Intelligent Education Systems (AAAI 2024 · 13 citations)
- 2025 · Generating Counterfactual Explanations Under Temporal Constraints (AAAI 2025 · 8 citations)
- 2026 · Autonomous Concept Drift Threshold Determination (AAAI 2026 · 1 citation)
Earliest papers here
- Collaborative Company Profiling: Insights from an Employee's Perspective (AAAI 2017 · 33 citations)
- Strategic Sequences of Arguments for Persuasion Using Decision Trees (AAAI 2017 · 23 citations)
- An Ambiguity Aversion Model for Decision Making under Ambiguity (AAAI 2017 · 4 citations)
What has been tried
64 papers here report 526 measured changes (a component added, removed or swapped, compared with a baseline): helped 70%, no clear effect 8%, hurt 17%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| add GP to learn and update the interaction function | Math1 · accuracy | 67.22 -> 67.78 | Helped | Symbolic Cognitive Diagnosis via Hybrid Optimization for Intelligent Education Systems (AAAI 2024 · 13 citations) |
| add the energy constraint to ECCCo (rather than remove the EBM term) | California Housing · Unfaithfulness | 1.70 ± 0.09 -> 1.40 ± 0.08 | Helped | Faithful Model Explanations through Energy-Constrained Conformal Counterfactuals (AAAI 2024 · 10 citations) |
| Add physical constraints to SimFair | AT1 West-East and East-Alaska · Fairness | 2.04 (±0.19) -> 1.89 (±0.06); 4.25 (±0.78) -> 4.45 (±0.51) | Hurt | SimFair: Physics-Guided Fairness-Aware Learning with Simulation Models (AAAI 2024 · 5 citations) |
| use causal intervention P(Y|do(X)) instead of self-attention computation | DS1 and DS2 lesion assessment · performance | Helped | A Trusted Lesion-assessment Network for Interpretable Diagnosis of Coronary Artery Disease in Coronary CT Angiography (AAAI 2025 · 3 citations) | |
| Add slipping modeling | ASSIST09 · ACC | 0.7140 -> 0.7367 | Helped | AD4CD: Causal-Guided Anomaly Detection for Enhancing Cognitive Diagnosis (AAAI 2025 · 2 citations) |
| Add JAC guidance to the creative ranker | Creative ranking test set · NSCTR | 0.2392 -> 0.2423 | Helped | Parallel Ranking of Ads and Creatives in Real-Time Advertising Systems (AAAI 2024 · 2 citations) |
| replace MHCN with current-state single-station MLP encoding | Nationwide terminal stations · MAE | 0.010±0.003 -> 0.026±0.002 | Hurt | Adaptive Multi-Faceted Service Capabilities Co-Prediction for Nationwide Terminal Stations in Logistics (AAAI 2025 · 1 citation) |
| Combine the amortized explanation with a Monte Carlo explanation using the optimized initial-guess weight instead of using Monte Carlo alone | UCI-Adult, UCI-News, Yelp Review, and Toxigen · MSE | Helped | Selective Explanations (NeurIPS 2024 · 1 citation) | |
| Use PEM instead of reverse diffusion for sampling | N-Queens · sample quality | Helped | Generalizable Reasoning through Compositional Energy Minimization (NeurIPS 2025 · 0 citations) | |
| Add information injection to LeapFactual | Morpho-MNIST · ACC | 0.9868±0.0026 -> 0.9906±0.0016 | Helped | LeapFactual: Reliable Visual Counterfactual Explanation Using Conditional Flow Matching (NeurIPS 2025 · 0 citations) |
| Increase the competition-cost weight λ₂ | Simulation varying λ₂ from 0.01 to 0.3 · modification cost | Mixed | Collective Counterfactual Explanations: Balancing Individual Goals and Collective Dynamics (NeurIPS 2025 · 0 citations) | |
| Set sparsity parameter p=0.01 rather than p=2.0 | Adult · continuous proximity | 0.581 -> 0.373 | Helped | DiCoFlex: Model-Agnostic Diverse Counterfactuals with Flexible Control (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 16 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- Adult · 2 of 16 papers
- Bank Marketing · 2 of 16 papers
- Lending Club · 2 of 16 papers
Baselines
- C-CHVAE · 3 of 16 papers
- Wachter · 2 of 16 papers
- DiCE · 2 of 16 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 55 papers; from OpenAlex and the papers' affiliations).
- Harvard University · 3 papers
- JD.com · 2 papers
- Free University of Bozen-Bolzano · 2 papers
- Delft University of Technology · 2 papers
- University of Pennsylvania · 2 papers
- Peking University · 2 papers
- Rutgers University · 2 papers
- Ludwig-Maximilians-Universität München · 2 papers
- Technion - Israel Institute of Technology · 2 papers
- Jagiellonian University · 2 papers
Use it
Download BibTeX of 37 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
- Prediction-powered and conformal inference (129 papers)
- Statistical analysis of ML practice (122 papers)
- Machine learning for physical simulation (96 papers)
- Sparse concepts and hypothesis testing (78 papers)
- Cold-start recommendation models (80 papers)
Elsewhere in the topic
- Generative protein structure design (142 papers)
- Neuro-symbolic probabilistic layers (120 papers)
- Structure-based 3D molecule generation (107 papers)
- Sparse graphical and tensor models (97 papers)
- Flow matching generative modeling (97 papers)
- Multimodal trajectory prediction (95 papers)
- Probabilistic embeddings and relational inference (93 papers)
- Equivariant and disentangled latent representations (90 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).