Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Interactive multi-agent trajectory prediction
Predicts joint future trajectories with relational reasoning, lane awareness, and factorized interaction models.
75 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Human-AI and goal-conditioned, keyword cluster “navigation · prediction · embodied”
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.
20 papers in 2023–24 and 17 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 6.8% to 4.4%.
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 · DESIRE: Distant Future Prediction in Dynamic Scenes With Interacting Agents (CVPR 2017 · 1,086 citations)
- 2019 · SoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints (CVPR 2019 · 1,115 citations)
- 2020 · VectorNet: Encoding HD Maps and Agent Dynamics From Vectorized Representation (CVPR 2020 · 967 citations)
- 2021 · LaPred: Lane-Aware Prediction of Multi-Modal Future Trajectories of Dynamic Agents (CVPR 2021 · 122 citations)
- 2022 · HiVT: Hierarchical Vector Transformer for Multi-Agent Motion Prediction (CVPR 2022 · 457 citations)
- 2023 · Query-Centric Trajectory Prediction (CVPR 2023 · 310 citations)
- 2024 · HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention (CVPR 2024 · 88 citations)
- 2025 · TraF-Align: Trajectory-aware Feature Alignment for Asynchronous Multi-agent Perception (CVPR 2025 · 10 citations)
- 2026 · Differentiable Semantic Meta-Learning Framework for Long-Tail Motion Forecasting in Autonomous Driving (AAAI 2026 · 2 citations)
Earliest papers here
- DESIRE: Distant Future Prediction in Dynamic Scenes With Interacting Agents (CVPR 2017 · 1,086 citations)
- Forecasting Interactive Dynamics of Pedestrians With Fictitious Play (CVPR 2017 · 155 citations)
- Agent-Centric Risk Assessment: Accident Anticipation and Risky Region Localization (CVPR 2017 · 72 citations)
What has been tried
65 papers here report 723 measured changes (a component added, removed or swapped, compared with a baseline): helped 75%, no clear effect 8%, hurt 15%, 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 the full Triple Factorized Attention module | Argoverse validation · b-minFDE | 1.832 -> 1.506 | Helped | HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention (CVPR 2024 · 88 citations) |
| Add velocity component | SDD · ADE | Helped | SocialCircle: Learning the Angle-based Social Interaction Representation for Pedestrian Trajectory Prediction (CVPR 2024 · 49 citations) | |
| increase anchor count from 1 to 2 | Argoverse val · minFDE | 0.928 -> 0.911 | Helped | SmartRefine: A Scenario-Adaptive Refinement Framework for Efficient Motion Prediction (CVPR 2024 · 41 citations) |
| Add local pose embedding | JRDB-GMP · JPE @ 5s | 471.4 -> 400.5 | Helped | Multi-agent Long-term 3D Human Pose Forecasting via Interaction-aware Trajectory Conditioning (CVPR 2024 · 19 citations) |
| replace the one-query-one-trajectory approach with decoupled mode and state queries | Argoverse 2 single-agent validation set · minFDE6 | 1.414 -> 1.268 | Helped | DeMo: Decoupling Motion Forecasting into Directional Intentions and Dynamic States (NeurIPS 2024 · 18 citations) |
| add field predictor | DAIR-V2X-Seq, infrastructure latency 0 ms · AP50/AP70 | 75.35/54.72 -> 76.31/57.02 | Helped | TraF-Align: Trajectory-aware Feature Alignment for Asynchronous Multi-agent Perception (CVPR 2025 · 10 citations) |
| add loss augmentation (λ > 0) | Basketball-U · minADE20[60] | 4.68 -> 4.65 | Helped | Unified Uncertainty-Aware Diffusion for Multi-Agent Trajectory Modeling (CVPR 2025 · 10 citations) |
| Add the motif matrix, ASI, and ATI instead of using the adjacency-matrix interaction baseline | nuScenes · ADE1 | 3.05 -> 2.84 | Helped | Density-Adaptive Model Based on Motif Matrix for Multi-Agent Trajectory Prediction (CVPR 2024 · 10 citations) |
| Add neighbor velocities (V) | SDD · ADE | 0.46 -> 0.45 | Helped | SocialMOIF: Multi-Order Intention Fusion for Pedestrian Trajectory Prediction (CVPR 2025 · 9 citations) |
| Increase context from 10s to 60s | Soccer trajectory reconstruction · mADE (players) | 4.13 -> 3.22 | Helped | Event2Tracking: Reconstructing Multi-Agent Soccer Trajectories Using Long-Term Multimodal Context (AAAI 2025 · 4 citations) |
| Add the profile module | split_244 · accuracy | 78.6 ± 0.10 -> 91.8 ± 0.15 | Helped | Political Actor Agent: Simulating Legislative System for Roll Call Votes Prediction with Large Language Models (AAAI 2025 · 2 citations) |
| Increase q+1 from 1 to 1.5 | Data centers: similar agents, different λ · Variance | 0.0029 -> 0.0002 | Helped | Building Socially-Equitable Public Models (ICML 2024 · 1 citation) |
What reviewers will expect
Datasets and baselines named in the 12 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- nuScenes · 2 of 12 papers
- ETH-UCY · 2 of 12 papers
- SDD · 2 of 12 papers
Baselines
- AutoBots · 2 of 12 papers
- MID · 2 of 12 papers
- SSSD · 2 of 12 papers
- GC-VRNN · 2 of 12 papers
- Naomi · 2 of 12 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 37 papers; from OpenAlex and the papers' affiliations).
- National University of Singapore · 3 papers
- Carnegie Mellon University · 2 papers
- Purdue University · 2 papers
- Fudan University · 2 papers
- Tsinghua University · 2 papers
- Peking University · 2 papers
- Korea Advanced Institute of Science and Technology · 2 papers
- University of Cambridge · 2 papers
Use it
Download BibTeX of 39 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
- Embodied agent learning (93 papers)
- Vision-language navigation agents (66 papers)
- Data-driven traffic simulation (36 papers)
Elsewhere in the topic
- Human-AI coordination agents (125 papers)
- World-model-based agents (110 papers)
- Imitation and skill learning (100 papers)
- Language-grounded reinforcement agents (96 papers)
- Boundedly rational agent modeling (87 papers)
- Collaborative multi-agent perception (86 papers)
- Intrinsically motivated exploration (86 papers)
- Learning-guided MAPF search (73 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).