Research map / Probabilistic and scientific modeling / Molecules and proteins
Sparse concepts and hypothesis testing
Defines and tests emergent sparse interactions and statistical significance in networks.
78 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.
27 papers in 2023–24 and 25 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 5.5% to 2.7%.
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 · Physics Inspired Optimization on Semantic Transfer Features: An Alternative Method for Room Layout Estimation (CVPR 2017 · 80 citations)
- 2018 · Variable Selection via Penalized Neural Network: a Drop-Out-One Loss Approach (ICML 2018 · 13 citations)
- 2019 · KnockoffGAN: Generating Knockoffs for Feature Selection using Generative Adversarial Networks (ICLR 2019 · 38 citations)
- 2020 · Deep active inference agents using Monte-Carlo methods (NeurIPS 2020 · 26 citations)
- 2021 · Explaining heterogeneity in medial entorhinal cortex with task-driven neural networks (NeurIPS 2021 · 38 citations)
- 2022 · Quantifying Statistical Significance of Neural Network-based Image Segmentation by Selective Inference (NeurIPS 2022 · 6 citations)
- 2023 · Physics-Guided ISO-Dependent Sensor Noise Modeling for Extreme Low-Light Photography (CVPR 2023 · 36 citations)
- 2024 · From GARCH to Neural Network for Volatility Forecast (AAAI 2024 · 13 citations)
- 2025 · Relational Neurosymbolic Markov Models (AAAI 2025 · 6 citations)
- 2026 · ECSEL: Explainable Classification via Signomial Equation Learning (ICML 2026 · 0 citations)
Earliest papers here
- Physics Inspired Optimization on Semantic Transfer Features: An Alternative Method for Room Layout Estimation (CVPR 2017 · 80 citations)
- NeuralFDR: Learning Discovery Thresholds from Hypothesis Features (NeurIPS 2017 · 12 citations)
- Variable Selection via Penalized Neural Network: a Drop-Out-One Loss Approach (ICML 2018 · 13 citations)
What has been tried
63 papers here report 582 measured changes (a component added, removed or swapped, compared with a baseline): helped 63%, no clear effect 12%, hurt 18%, mixed 7%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Replace MSE loss with maximum-likelihood N-loss | Multiple datasets · MAE | Helped | From GARCH to Neural Network for Volatility Forecast (AAAI 2024 · 13 citations) | |
| combine diffusion and advection processes | Beijing and Shenzhen, 24h–72h · MAE | Helped | AirPhyNet: Harnessing Physics-Guided Neural Networks for Air Quality Prediction (ICLR 2024 · 5 citations) | |
| Use minimum leaf size 20 | Proof of concept · type I error rate and power | 5%; 100% | Helped | Conditional Feature Importance with Generative Modeling Using Adversarial Random Forests (AAAI 2025 · 3 citations) |
| Add both 1-D and 2-D temporal predicates | Sentiment analysis · explanation fidelity | Helped | ReX: A Framework for Incorporating Temporal Information in Model-Agnostic Local Explanation Techniques (AAAI 2025 · 1 citation) | |
| vary the number of unknown parameters to 2 instead of 4 | Forced Damped Pendulum · MCC | 0.99(0.00) -> 0.97(0.01) | No clear effect | On the Identifiability of Hybrid Deep Generative Models: Meta-Learning as a Solution (NeurIPS 2024 · 1 citation) |
| add the non-shared deep architecture variant (b) | Air quality (NO2) · correlation | 0.880 (0.052) -> 0.897 (0.047) | No clear effect | A Functional Extension of Semi-Structured Networks (NeurIPS 2024 · 1 citation) |
| Use an implicit time-varying loss instead of an explicit loss | Synthetic time-varying potential · potential recovery | Helped | Learning diffusion at lightspeed (NeurIPS 2024 · 1 citation) | |
| Replace random sampling after Gradient-based Filtering with Multigrid Genetic Sampling | Four biological systems · RMSE | Helped | Hierarchical Gradient-Based Genetic Sampling for Accurate Prediction of Biological Oscillations (AAAI 2025 · 0 citations) | |
| use Top probability ordering instead of vanilla decoding | Sudoku · accuracy | 6.88% -> 18.51% | Helped | Train for the Worst, Plan for the Best: Understanding Token Ordering in Masked Diffusions (ICML 2025 · 0 citations) |
| Increase GradSHAP(t) integration samples | Simulated survival data · local accuracy | Helped | Gradient-based Explanations for Deep Learning Survival Models (ICML 2025 · 0 citations) | |
| replace Gaussian RBF with soft distillation | mESC · AUPRC | 0.646 -> 0.636 | Hurt | KINDLE: Knowledge-Guided Distillation for Prior-Free Gene Regulatory Network Inference (NeurIPS 2025 · 0 citations) |
| add bidirectional attention | Synthetic Case 1, N=200 · MSE | 0.0812 -> 0.0783 | Helped | FAME: Adaptive Functional Attention with Expert Routing for Function-on-Function Regression (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 17 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- KernelSHAP · 3 of 17 papers
- LIME · 2 of 17 papers
- Random · 2 of 17 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 52 papers; from OpenAlex and the papers' affiliations).
- Shanghai Jiao Tong University · 5 papers
- University of California, San Diego · 2 papers
- Northeastern University · 2 papers
- Peking University · 2 papers
- Harvard 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)
- Counterfactual and statistical explanations (81 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).