Research map / Graph learning / Knowledge and temporal graphs
Continuous-time dynamic graph learning
Model link prediction on continuous-time dynamic graphs and benchmark temporal graph learning.
132 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Knowledge and temporal graphs, keyword cluster “dynamic · temporal graph · spatio temporal”
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.
38 papers in 2023–24 and 73 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 14.7% to 25.8%.
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 · Unfolding Temporal Dynamics: Predicting Social Media Popularity Using Multi-scale Temporal Decomposition (AAAI 2016 · 113 citations)
- 2017 · EEG-GRAPH: A Factor-Graph-Based Model for Capturing Spatial, Temporal, and Observational Relationships in Electroencephalograms (NeurIPS 2017)
- 2019 · DyRep: Learning Representations over Dynamic Graphs (ICLR 2019 · 224 citations)
- 2020 · Disentangling and Unifying Graph Convolutions for Skeleton-Based Action Recognition (CVPR 2020 · 1,219 citations)
- 2021 · Transfer Graph Neural Networks for Pandemic Forecasting (AAAI 2021 · 154 citations)
- 2022 · DDGCN: Dual Dynamic Graph Convolutional Networks for Rumor Detection on Social Media (AAAI 2022 · 118 citations)
- 2023 · Uncovering the Missing Pattern: Unified Framework Towards Trajectory Imputation and Prediction (CVPR 2023 · 44 citations)
- 2024 · Temporal Graph Contrastive Learning for Sequential Recommendation (AAAI 2024 · 62 citations)
- 2025 · A Generalizable Anomaly Detection Method in Dynamic Graphs (AAAI 2025 · 15 citations)
- 2026 · Hyperbolic-Enhanced Mixture-of-Experts Mamba for Sequential Recommendation (AAAI 2026 · 3 citations)
Earliest papers here
- Unfolding Temporal Dynamics: Predicting Social Media Popularity Using Multi-scale Temporal Decomposition (AAAI 2016 · 113 citations)
- Reconstructing Parameters of Spreading Models from Partial Observations (NeurIPS 2016 · 22 citations)
- EEG-GRAPH: A Factor-Graph-Based Model for Capturing Spatial, Temporal, and Observational Relationships in Electroencephalograms (NeurIPS 2017)
What has been tried
84 papers here report 808 measured changes (a component added, removed or swapped, compared with a baseline): helped 76%, no clear effect 3%, 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 disturbed graph contrastive learning (DGCL) | Beauty and Comics · NDCG@10, NDCG@20 | Helped | Temporal Graph Contrastive Learning for Sequential Recommendation (AAAI 2024 · 62 citations) | |
| Add the node-level distribution module | three ablation-study datasets · clustering performance | Helped | Deep Temporal Graph Clustering (ICLR 2024 · 24 citations) | |
| add Uncertainty Calibration Module | Gowalla · Acc@1 | 0.1487 -> 0.1604 | Helped | TAU: Trajectory Data Augmentation with Uncertainty for Next POI Recommendation (AAAI 2024 · 23 citations) |
| add entity nodes to the graph | Qin’s and MAEC · stock movement prediction performance | Helped | ECHO-GL: Earnings Calls-Driven Heterogeneous Graph Learning for Stock Movement Prediction (AAAI 2024 · 19 citations) | |
| Integrate the geometric subspaces with HGE instead of stacking | ICEWS14, ICEWS05-15, and GDELT · MRR | 62.0 -> 63.0; 67.3 -> 68.1; 35.6 -> 37.1 | Helped | HGE: Embedding Temporal Knowledge Graphs in a Product Space of Heterogeneous Geometric Subspaces (AAAI 2024 · 18 citations) |
| Remove temporal ego-graph sampling | Bitcoin-Alpha · AP | 26.73 -> 19.33 | Helped | A Generalizable Anomaly Detection Method in Dynamic Graphs (AAAI 2025 · 15 citations) |
| Add DMI regularization | Tmall · HR@20 | 38.27 -> 40.26 | Helped | Dynamic Multi-Interest Graph Neural Network for Session-Based Recommendation (AAAI 2025 · 15 citations) |
| Add a trainable multi-view time encoder | D1 and D2 · AUC | Helped | Dynamic Graph Learning with Static Relations for Credit Risk Assessment (AAAI 2025 · 13 citations) | |
| add cross-temporal representation learning | DutchWind, BeijingMEO, LondonAQ, and CN, 10% missing · RMSE | 0.4221 -> 0.4101; 0.4308 -> 0.3986; 0.3515 -> 0.2720; 0.3723 -> 0.2534 | Helped | Graph Structure Learning for Spatial-Temporal Imputation: Adapting to Node and Feature Scales (AAAI 2025 · 11 citations) |
| add edge-type information to TGN | thgl-software · MRR | 0.324 -> 0.424 | Helped | TGB 2.0: A Benchmark for Learning on Temporal Knowledge Graphs and Heterogeneous Graphs (NeurIPS 2024 · 9 citations) |
| Add pairwise node representations and decoded pairwise features | Temporal link prediction · performance | Helped | Improving Temporal Link Prediction via Temporal Walk Matrix Projection (NeurIPS 2024 · 9 citations) | |
| Add BERT-encoded text initialization to dynamic graph models | Edge classification across DTGB datasets · Weighted F1 | Helped | DTGB: A Comprehensive Benchmark for Dynamic Text-Attributed Graphs (NeurIPS 2024 · 8 citations) |
What reviewers will expect
Datasets and baselines named in the 47 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- LastFM · 6 of 47 papers
- Reddit · 4 of 47 papers
- Wikipedia · 4 of 47 papers
- MOOC · 4 of 47 papers
- UN Vote · 4 of 47 papers
- Enron · 3 of 47 papers
- Social Evo. · 3 of 47 papers
- UCI · 3 of 47 papers
Baselines
- TGAT · 7 of 47 papers
- TGN · 6 of 47 papers
- STGCN · 6 of 47 papers
- JODIE · 5 of 47 papers
- DyRep · 5 of 47 papers
- CAWN · 5 of 47 papers
- GraphMixer · 5 of 47 papers
- DyGFormer · 4 of 47 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 111 papers; from OpenAlex and the papers' affiliations).
- Hong Kong University of Science and Technology · 7 papers
- Peking University · 4 papers
- University of Science and Technology of China · 4 papers
- Nanyang Technological University · 4 papers
- Tsinghua University · 4 papers
- Université de Montréal · 4 papers
- The University of Tokyo · 3 papers
- National University of Singapore · 3 papers
- University of Illinois, Urbana Champaign · 3 papers
- Tencent · 3 papers
Use it
Download BibTeX of 35 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
- Temporal graph neural architectures (69 papers)
- Spatio-temporal graph traffic forecasting (60 papers)
Elsewhere in the topic
- Knowledge graph embeddings (110 papers)
- Graph reasoning for visual recognition (96 papers)
- Graph-based dialog systems (94 papers)
- Social and behavior graphs (87 papers)
- Temporal knowledge graph reasoning (84 papers)
- Scene graph generation (70 papers)
- Logical query answering and rule reasoning (68 papers)
- Heterogeneous graph learning (66 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).