Research map / Probabilistic and scientific modeling / Molecules and proteins
Neural point-process models
Models event sequences with attention, latent processes and Hawkes or graphical event models.
64 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Molecules and proteins, keyword cluster “event · point processes · temporal”
State History Tried 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.
11 papers in 2023–24 and 11 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 2.2% to 1.2%.
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 · Isotonic Hawkes Processes (ICML 2016 · 47 citations)
- 2017 · The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process (NeurIPS 2017 · 204 citations)
- 2018 · Learning the Joint Representation of Heterogeneous Temporal Events for Clinical Endpoint Prediction (AAAI 2018 · 43 citations)
- 2019 · SAFE: A Neural Survival Analysis Model for Fraud Early Detection (AAAI 2019 · 40 citations)
- 2020 · Event Probability Mask (EPM) and Event Denoising Convolutional Neural Network (EDnCNN) for Neuromorphic Cameras (CVPR 2020 · 97 citations)
- 2021 · Clinical Temporal Relation Extraction with Probabilistic Soft Logic Regularization and Global Inference (AAAI 2021 · 47 citations)
- 2022 · Anytime-Valid Inference For Multinomial Count Data (NeurIPS 2022 · 6 citations)
- 2023 · Concurrent Multi-Label Prediction in Event Streams (AAAI 2023 · 8 citations)
- 2024 · A Variational Autoencoder for Neural Temporal Point Processes with Dynamic Latent Graphs (AAAI 2024 · 3 citations)
- 2025 · GeoMamba: Towards Multi-granular POI Recommendation with Geographical State Space Model (AAAI 2025 · 4 citations)
- 2026 · Marrying Generative Model of Healthcare Events with Digital Twin of Social Determinants of Health for Disease Reasoning (ICML 2026 · 0 citations)
Earliest papers here
- Isotonic Hawkes Processes (ICML 2016 · 47 citations)
- Bayesian Poisson Tucker Decomposition for Learning the Structure of International Relations (ICML 2016 · 41 citations)
- The Ostomachion Process (AAAI 2016 · 11 citations)
What has been tried
56 papers here report 490 measured changes (a component added, removed or swapped, compared with a baseline): helped 73%, no clear effect 7%, hurt 15%, 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 |
|---|---|---|---|---|
| use specialized encoders instead of vanilla transformer layers | Singapore · Recall@2 | 0.2719 -> 0.3740 | Helped | GeoMamba: Towards Multi-granular POI Recommendation with Geographical State Space Model (AAAI 2025 · 4 citations) |
| Use a dynamic rather than static latent graph | MathOF, AskUbuntu, SuperUser, NYMVC · negative log-likelihood | MathOF: -2.632±0.028 -> -3.501±0.068; AskUbuntu: -2.312±0.026 -> -2.867±0.032; SuperUser: -2.466±0.021 -> -3.812±0.057; | Helped | A Variational Autoencoder for Neural Temporal Point Processes with Dynamic Latent Graphs (AAAI 2024 · 3 citations) |
| Add integral normalization | HQF motion deblurring · MSE | 0.264 -> 0.152 | Helped | PPLNs: Parametric Piecewise Linear Networks for Event-Based Temporal Modeling and Beyond (NeurIPS 2024 · 2 citations) |
| Add the Node Update stage | Enron · loss | 5.1% | Helped | Deep Representation Learning for Forecasting Recursive and Multi-Relational Events in Temporal Networks (AAAI 2025 · 1 citation) |
| Set τ to 10.0 rather than 0.1 | Synthetic, I=1000, J=100, σ=0.5 · Kendall's tau | 0.01 -> 0.81 | Helped | Unscrambling disease progression at scale: fast inference of event permutations with optimal transport (NeurIPS 2024 · 1 citation) |
| Use Bi-CMEs instead of separately learned permutation and magnitude CMEs | West Virginia-0 · reduced intensity | 0.01(0.06) -> 0.54(0.27); 0.05(0.02) -> 0.54(0.27) | Helped | Amortized Network Intervention to Steer the Excitatory Point Processes (ICLR 2024 · 0 citations) |
| Combine sarr and sint instead of using either score alone | GOF test for SPP: Stopping and RenewalA · AUROC | Helped | Conformal Anomaly Detection in Event Sequences (ICML 2025) | |
| add prior knowledge | Syn Data-1 · ER% | 41.72 -> 45.64 | Helped | Evolving Minds: Logic-Informed Inference from Temporal Action Patterns (ICML 2025) |
| include a coordination variable | Synthetic D_c · RMSE | 0.18 (.004) -> 0.11 (.004) | Helped | Probabilistic Modeling of Interpersonal Coordination Processes (ICML 2024) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 22 papers; from OpenAlex and the papers' affiliations).
- IBM · 3 papers
- Rensselaer Polytechnic Institute · 3 papers
- The Chinese University of Hong Kong · 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
- Survival analysis under censoring (12 papers)
Elsewhere in the topic
- Generative protein structure design (142 papers)
- Prediction-powered and conformal inference (129 papers)
- Statistical analysis of ML practice (122 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)
- Machine learning for physical simulation (96 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).