Research map / Multi-agent systems and games / Human-AI and goal-conditioned
EqMotion: Equivariant Multi-Agent Motion Prediction With Invariant Interaction Reasoning
Chenxin Xu, Robby T. Tan, Yuhong Tan, Siheng Chen, Yu Guang Wang, Xinchao Wang, Yanfeng Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use EqMotion's equivariant geometric, invariant pattern, and invariant interaction reasoning modules | motion prediction · MD17 molecules | ADE/FDE: 34.2%/30.1% | Improved |
| Use EqMotion | H3.6M short-term skeleton prediction · H3.6M | MPJPE: 8.6% | Improved |
| Use EqMotion | H3.6M long-term skeleton prediction · H3.6M | MPJPE: 3.8% | Improved |
| Use EqMotion | deterministic pedestrian prediction · ETH-UCY | FDE: 10.4% | Improved |
| Use EqMotion multi-prediction | pedestrian prediction · ETH-UCY | FDE: 7.9% | Improved |
| Add EGFL (ablation) | EqMotion · H3.6M | MPJPE at 400ms: 82.4 -> 60.7 | Improved |
| Add IPFL (ablation) | EqMotion with EGFL · H3.6M | MPJPE at 400ms: 60.7 -> 57.0 | Improved |
| Add IRM (ablation) | EqMotion with EGFL and IPFL · H3.6M | MPJPE at 400ms: 57.0 -> 55.0 | Improved |
| Add inner-agent attention (ablation) | EqMotion · H3.6M | MPJPE at 400ms: 55.7 -> 55.0 | Improved |
| Add inter-agent aggregation (ablation) | EqMotion · H3.6M | MPJPE at 400ms: 58.9 -> 55.0 | Improved |
| Add equivariant nonlinear function (ablation) | EqMotion · H3.6M | MPJPE at 400ms: 58.6 -> 55.0 | Improved |
| Increase feature-learning depth from 1 to 4 layers (ablation) | EqMotion · H3.6M | MPJPE at 400ms: 58.3 -> 55.0 | Improved |
Later papers that compared with this method
1 later accepted paper reports EqMotion as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- MART: MultiscAle Relational Transformer Networks for Multi-agent Trajectory Prediction (ECCV 2024 · better on 3 of 3 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- EvolveGraph: Multi-Agent Trajectory Prediction with Dynamic Relational Reasoning (NeurIPS 2020 · similarity 0.86)
- Equivariant Graph Neural Operator for Modeling 3D Dynamics (ICML 2024 · similarity 0.85)
- Online Relational Inference for Evolving Multi-agent Interacting Systems (NeurIPS 2024 · similarity 0.84)
- Deep SE(3)-Equivariant Geometric Reasoning for Precise Placement Tasks (ICLR 2024 · similarity 0.84)
- Large Language-Geometry Model: When LLM meets Equivariance (ICML 2025 · similarity 0.84)
- Hierarchical Equivariant Policy via Frame Transfer (ICML 2025 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
