Research map / Graph learning / Knowledge and temporal graphs
Social and behavior graphs
Model social interactions with graphs for trajectory and behavior prediction.
87 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Knowledge and temporal graphs, keyword cluster “scene · image · visual”
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.
26 papers in 2023–24 and 35 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 10.1% to 12.4%.
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 · Leveraging Node Attributes for Incomplete Relational Data (ICML 2017 · 19 citations)
- 2018 · Ranking Users in Social Networks With Higher-Order Structures (AAAI 2018 · 53 citations)
- 2019 · Session-Based Recommendation with Graph Neural Networks (AAAI 2019 · 1,530 citations)
- 2020 · Recursive Social Behavior Graph for Trajectory Prediction (CVPR 2020 · 216 citations)
- 2021 · Knowledge-Enhanced Hierarchical Graph Transformer Network for Multi-Behavior Recommendation (AAAI 2021 · 234 citations)
- 2022 · Heterogeneity-Aware Twitter Bot Detection with Relational Graph Transformers (AAAI 2022 · 117 citations)
- 2023 · Dynamic Graph Enhanced Contrastive Learning for Chest X-Ray Report Generation (CVPR 2023 · 188 citations)
- 2024 · LGMRec: Local and Global Graph Learning for Multimodal Recommendation (AAAI 2024 · 146 citations)
- 2025 · PRAGA: Prototype-aware Graph Adaptive Aggregation for Spatial Multi-modal Omics Analysis (AAAI 2025 · 18 citations)
- 2026 · Hierarchical Attention Network with Correction for Cross-Domain User Association (AAAI 2026 · 2 citations)
Earliest papers here
- Leveraging Node Attributes for Incomplete Relational Data (ICML 2017 · 19 citations)
- Ranking Users in Social Networks With Higher-Order Structures (AAAI 2018 · 53 citations)
- Session-Based Recommendation with Graph Neural Networks (AAAI 2019 · 1,530 citations)
What has been tried
66 papers here report 700 measured changes (a component added, removed or swapped, compared with a baseline): helped 81%, no clear effect 3%, hurt 13%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add multimodal contents | Baby, Sports, Clothing · Recall@20 | 0.0732 -> 0.1002; 0.0829 -> 0.1068; 0.0514 -> 0.0828 | Helped | LGMRec: Local and Global Graph Learning for Multimodal Recommendation (AAAI 2024 · 146 citations) |
| Construct directed edges using head-tail similarity rather than k-NN cosine similarity | TCGA-ESCA typing · AUC | 94.982.52 -> 95.232.90 | Helped | Dynamic Graph Representation with Knowledge-aware Attention for Histopathology Whole Slide Image Analysis (CVPR 2024 · 94 citations) |
| Combine the molecular graph and biomedical knowledge graph channels | DrugBank · ACC | 0.6347 -> 0.7903; 0.7126 -> 0.7903 | Helped | Dual-Channel Learning Framework for Drug-Drug Interaction Prediction via Relation-Aware Heterogeneous Graph Transformer (AAAI 2024 · 50 citations) |
| use a collision-aware kernel with distance scaling rather than no weights | ETH/UCY AVG · minADE20 | 0.49 -> 0.39 | Helped | Higher-order Relational Reasoning for Pedestrian Trajectory Prediction (CVPR 2024 · 31 citations) |
| Replace the learnable omics-specific graph with a KNN graph | Human Lymph Node · NMI | 39.47 -> 38.69 | Hurt | PRAGA: Prototype-aware Graph Adaptive Aggregation for Spatial Multi-modal Omics Analysis (AAAI 2025 · 18 citations) |
| Add TME neighborhood organization guided graph embedding instead of omitting it or using DeepWalk | BRCA, KIRC, LUSC · CI | TMEGL: 0.719, 0.697, 0.695; TMEGL-NE: 0.609, 0.629, 0.618; TMEGL-DW: 0.658, 0.631, 0.621 | Helped | Tumor Micro-environment Interactions Guided Graph Learning for Survival Analysis of Human Cancers from Whole-slide Pathological Images (CVPR 2024 · 14 citations) |
| Add feature completion modeling | MovieLens100K UCS; MovieLens1M UCS; Yelp UCS; MovieLens100K ICS; MovieLens1M ICS; Yelp ICS · MAE | 0.8359 -> 0.7735; 0.8427 -> 0.7653; 0.9305 -> 0.8492; 0.8805 -> 0.8064; 0.8711 -> 0.7671; 0.8967 -> 0.8002 | Helped | Feature-Structure Adaptive Completion Graph Neural Network for Cold-start Recommendation (AAAI 2025 · 10 citations) |
| use the multi-faceted hierarchical graph instead of all 1-hop neighbors | ST-Net · PCC | 0.5789 -> 0.6795 | Helped | MERGE: Multi-faceted Hierarchical Graph-based GNN for Gene Expression Prediction from Whole Slide Histopathology Images (CVPR 2025 · 10 citations) |
| add Stage 1 latent network inference | FakeSV · accuracy | 81.68 -> 89.14 | Helped | Learning Complex Heterogeneous Multimodal Fake News via Social Latent Network Inference (AAAI 2025 · 8 citations) |
| Add CorrGCN | MAMS · accuracy | 84.06 -> 86.08 | Helped | You Only Read Once: Constituency-Oriented Relational Graph Convolutional Network for Multi-Aspect Multi-Sentiment Classification (AAAI 2024 · 8 citations) |
| Use linear attention instead of Conv1D for item capsule projection | HV-E · Recall@5 | 58.33 -> 61.35 | Helped | Lightweight Yet Fine-Grained: A Graph Capsule Convolutional Network with Subspace Alignment for Shared-Account Sequential Recommendation (AAAI 2025 · 7 citations) |
| add spatial links to the heterogeneous graphs | DARTE 2015 and 2017 · R^2 | 0.652 -> 0.750; 0.719 -> 0.787 | Helped | Estimating On-Road Transportation Carbon Emissions from Open Data of Road Network and Origin-Destination Flow Data (AAAI 2024 · 7 citations) |
What reviewers will expect
Datasets and baselines named in the 19 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- DLPFC · 2 of 19 papers
Baselines
- LightGCN · 3 of 19 papers
- BERT · 2 of 19 papers
- STAGATE · 2 of 19 papers
- BLEEP · 2 of 19 papers
- HisToGene · 2 of 19 papers
- Hist2ST · 2 of 19 papers
- THItoGene · 2 of 19 papers
- TRIPLEX · 2 of 19 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 61 papers; from OpenAlex and the papers' affiliations).
- The University of Hong Kong · 4 papers
- University of Chinese Academy of Sciences · 4 papers
- Tsinghua University · 4 papers
- Zhejiang University · 3 papers
- Beijing University of Posts and Telecommunications · 2 papers
- University of Electronic Science and Technology of China · 2 papers
- University of Cambridge · 2 papers
- Sun Yat-sen University · 2 papers
- Tianjin University · 2 papers
- Chinese Academy of Sciences · 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
- Graph reasoning for visual recognition (96 papers)
- Graph-based dialog systems (94 papers)
- Heterogeneous graph learning (66 papers)
- Scene graph generation (70 papers)
Elsewhere in the topic
- Continuous-time dynamic graph learning (132 papers)
- Knowledge graph embeddings (110 papers)
- Temporal knowledge graph reasoning (84 papers)
- Temporal graph neural architectures (69 papers)
- Logical query answering and rule reasoning (68 papers)
- Spatio-temporal graph traffic forecasting (60 papers)
- Inductive knowledge graph completion (38 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).