Research map / Time series / Causal and generative time series
Neural differential equations
Model irregular time series with neural ODEs, SDEs and controlled differential equations.
52 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Causal and generative time series, keyword cluster “differential · continuous · equations”
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.
16 papers in 2023–24 and 18 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 10.3% to 7.9%.
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 · Solving Temporal Puzzles (CVPR 2016 · 5 citations)
- 2019 · Latent Ordinary Differential Equations for Irregularly-Sampled Time Series (NeurIPS 2019 · 238 citations)
- 2020 · Neural Controlled Differential Equations for Irregular Time Series (NeurIPS 2020 · 153 citations)
- 2021 · Liquid Time-constant Networks (AAAI 2021 · 309 citations)
- 2022 · Modeling Irregular Time Series with Continuous Recurrent Units (ICML 2022 · 8 citations)
- 2023 · ContiFormer: Continuous-Time Transformer for Irregular Time Series Modeling (NeurIPS 2023 · 27 citations)
- 2024 · IVP-VAE: Modeling EHR Time Series with Initial Value Problem Solvers (AAAI 2024 · 10 citations)
- 2025 · DualDynamics: Synergizing Implicit and Explicit Methods for Robust Irregular Time Series Analysis (AAAI 2025 · 7 citations)
- 2026 · Continuum Dropout for Neural Differential Equations (AAAI 2026 · 1 citation)
Earliest papers here
- Solving Temporal Puzzles (CVPR 2016 · 5 citations)
- Latent Ordinary Differential Equations for Irregularly-Sampled Time Series (NeurIPS 2019 · 238 citations)
- GRU-ODE-Bayes: Continuous Modeling of Sporadically-Observed Time Series (NeurIPS 2019 · 167 citations)
What has been tried
40 papers here report 392 measured changes (a component added, removed or swapped, compared with a baseline): helped 67%, no clear effect 8%, hurt 21%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use the flow output ẑ(t) instead of z(t) | BasicMotions, 70% Missing · Accuracy | 0.883 (0.095) -> 0.983 (0.037) | Helped | DualDynamics: Synergizing Implicit and Explicit Methods for Robust Irregular Time Series Analysis (AAAI 2025 · 7 citations) |
| Add a controlled path to the drift | UEA/UCR 30 datasets, all datasets · accuracy | 0.517 (0.064) -> 0.717 (0.051) | Helped | Stable Neural Stochastic Differential Equations in Analyzing Irregular Time Series Data (ICLR 2024 · 6 citations) |
| remove uncertainty prediction | Clinical trajectory modeling · Mean MSE (× 10^-3) | 0.793 ± 0.017 -> 1.170 ± 0.014; 2.762 ± 0.017 -> 2.759 ± 0.015; 2.673 ± 0.069 -> 3.097 ± 0.054; 8.245 ± 0.495 -> 8.659 ± | Hurt | Trajectory Flow Matching with Applications to Clinical Time Series Modelling (NeurIPS 2024 · 3 citations) |
| Learn Fourier mapping parameters rather than fixing them at initialization | Electricity interpolation, 1 missing block · MSE | 0.0079 -> 0.0061 | Helped | Neural Functions for Learning Periodic Signal (ICLR 2025 · 0 citations) |
| use stabilized weights instead of unstabilized weights | MIMIC-III, 1 hour · RMSE | 0.876± 0.041 -> 0.877±0.044 | No clear effect | Stabilized Neural Prediction of Potential Outcomes in Continuous Time (ICLR 2025 · 0 citations) |
| add the first observed time-series value to the static features | OU · IBS | 0.091±0.01 -> 0.095±0.01 | Hurt | Dynamic Survival Analysis with Controlled Latent States (ICML 2024 · 0 citations) |
| Set latent dimension m to 10 vs larger values (10 to 50) for parameter Koopman observables | Non-linear Oscillator · test MSE | Mixed | KoNODE: Koopman-Driven Neural Ordinary Differential Equations with Evolving Parameters for Time Series Analysis (ICML 2025) |
What reviewers will expect
Datasets and baselines named in the 9 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- ETTH1 · 2 of 9 papers
- ETTH2 · 2 of 9 papers
- Electricity · 2 of 9 papers
- Traffic · 2 of 9 papers
Baselines
- LSTM · 2 of 9 papers
- Transformer · 2 of 9 papers
- Latent-ODE · 2 of 9 papers
- PatchTST · 2 of 9 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 34 papers; from OpenAlex and the papers' affiliations).
- University of Oxford · 5 papers
- Ulsan National Institute of Science and Technology · 2 papers
- Yale University · 2 papers
- Inria · 2 papers
- Imperial College London · 2 papers
- Fudan University · 2 papers
Use it
Download BibTeX of 36 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
- Irregular series and imputation (59 papers)
- Diffusion models for time series (55 papers)
Elsewhere in the topic
- Probabilistic and conformal forecasting (121 papers)
- State-space and latent dynamics forecasting (94 papers)
- Robust forecasting and evaluation (91 papers)
- Dynamical system reconstruction (90 papers)
- Generative temporal point processes (53 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).