Research map / Multi-agent systems and games / Human-AI and goal-conditioned
SRefiner: Soft-Braid Attention for Multi-Agent Trajectory Refinement
Liwen Xiao, Zhiyu Pan, Zhicheng Wang, Zhiguo Cao, Wei Li
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 |
|---|---|---|---|
| add Traj-Traj soft-braid attention (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.514 -> 1.477 | Improved |
| add Traj-Lane soft-braid attention (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.497 -> 1.477 | Improved |
| add topology update across iterations (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.522 -> 1.477 | Improved |
| add soft-braid topology instead of no topology (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.530 -> 1.477 | Improved |
| use soft-braid topology instead of braid topology (BeTop) (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.512 -> 1.477 | Improved |
| add Traj-Lane topology to Traj-Traj soft-braid (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.497 -> 1.477 | Improved |
| add topology update (FJMP baseline) (ablation) | SRefiner on FJMP · Argoverse v2 val (FJMP baseline, ablation) | avgMinFDE: 1.787 -> 1.736 | Improved |
| add soft-braid attention vs no topology on INTERACTIONS (FJMP) (ablation) | SRefiner on FJMP · INTERACTIONS val (FJMP baseline, ablation) | minJointFDE: 0.601 -> 0.579 | Improved |
| vary T-T local radius from 10 to 50 (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.489 -> 1.477 | No clear effect |
| increase T-T local radius from 50 to 100 (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.477 -> 1.480 | No clear effect |
| vary T-L local radius from 2 to 10 (ablation) | SRefiner · Argoverse v2 val (Forecast-MAE baseline, ablation) | avgMinFDE: 1.484 -> 1.477 | No clear effect |
| apply SRefiner to HPNet (strong baseline) | SRefiner · INTERACTIONS test | minJointFDE: 0.823 -> 0.797 | Improved |
Later papers that compared with this method
1 later accepted paper reports SRefiner as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- SHARP: Short-Window Streaming for Accurate and Robust Prediction in Motion Forecasting (CVPR 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- FJMP — FJMP: Factorized Joint Multi-Agent Motion Prediction Over Learned Directed Acyclic Interaction Graphs (CVPR 2023) · better on 4 of 4 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
- DeMo: Decoupling Motion Forecasting into Directional Intentions and Dynamic States (NeurIPS 2024 · similarity 0.87)
- SmartRefine: A Scenario-Adaptive Refinement Framework for Efficient Motion Prediction (CVPR 2024 · similarity 0.86)
- Density-Adaptive Model Based on Motif Matrix for Multi-Agent Trajectory Prediction (CVPR 2024 · similarity 0.86)
- Trajeglish: Traffic Modeling as Next-Token Prediction (ICLR 2024 · similarity 0.86)
- HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention (CVPR 2024 · similarity 0.86)
- Reasoning Multi-Agent Behavioral Topology for Interactive Autonomous Driving (NeurIPS 2024 · similarity 0.86)
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.
