Research map / Multi-agent systems and games / Human-AI and goal-conditioned
The Role of Data-Driven Priors in Multi-Agent Crowd Trajectory Estimation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 a GP global flow prior rather than the NN-only prior | Trajectory interpolation framework · Basic evaluation across six scenarios | DTW distance: 3.67 vs 8.33 average rank | Helped |
| Add a local NN collision-avoidance prior alongside GP | Trajectory interpolation framework · Basic evaluation across six scenarios | Agent-agent collisions: 3.50 vs 5.67 average rank | Helped |
| Use GP alone without an additional collision policy learner | Trajectory interpolation framework · Basic evaluation across six scenarios | Agent-agent collisions: 4.41 vs 3.50 average rank | Hurt |
| Use GP-fed-ORCA instead of GP or NN-based collision policies | Trajectory interpolation framework · Basic evaluation across six scenarios | Agent-agent collisions: 7.58 vs 4.41 average rank | Hurt |
| Use GP global flow information in the prior | Trajectory interpolation framework · Basic evaluation across six scenarios | DTW distance: 3.67 vs 8.33 average rank | Helped |
| Use an optimizer that avoids pairwise collision constraints | Proposed interpolation framework · Basic evaluation | computational time | Helped |
| Choose IPM instead of UKS | Proposed interpolation framework · Basic evaluation across six scenarios | DTW distance | Helped |
| Choose UKS instead of IPM | Proposed interpolation framework · Basic evaluation across six scenarios | agent-agent collisions | Helped |
| Use MPA instead of IPM or UKS | Trajectory interpolation framework · Basic evaluation across six scenarios | agent-obstacle collisions | Helped |
| Test at densities different from the training density | Proposed prior methods · Extended evaluation | agent-agent collisions | Mixed |
| Use a local NN policy in the symmetric concentric-circle scenario | Trajectory interpolation framework · Concentric-circles extended evaluation | DTW distance | Helped |
| Increase agent density | Prior methods · Extended evaluation across scenarios | collisions | Hurt |
Closest papers by meaning
- Factored Online Planning in Many-Agent POMDPs (AAAI 2024 · similarity 0.86)
- Simulation-Informed Diffusion for Decentralized Multi-robot Motion Planning (NeurIPS 2026 · similarity 0.84)
- ATLAS: Constraints-Aware Multi-Agent Collaboration for Real-World Travel Planning (ICLR 2026 · similarity 0.83)
- Adaptive Anytime Multi-Agent Path Finding Using Bandit-Based Large Neighborhood Search (AAAI 2024 · similarity 0.83)
- HiMAP-Travel: Hierarchical Multi-Agent Planning for Long-Horizon Constrained Travel (ICML 2026 · similarity 0.83)
- Learn to Follow: Decentralized Lifelong Multi-Agent Pathfinding via Planning and Learning (AAAI 2024 · 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.