Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Interaction Modeling with Multiplex Attention
Fan-Yun Sun, Isaac Kauvar, Ruohan Zhang, Jiachen Li, Mykel Kochenderfer, Jiajun Wu, Nick Haber
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 multiplex attentional latent graph instead of single edge-type graph (ablation) | IMMA · Social Navigation | ADE: 0.156 -> 0.148 | Improved |
| add Progressive Layer Training (ablation) | IMMA (MG) · Social Navigation | ADE: 0.148 -> 0.139 | Improved |
| add Progressive Layer Training (ablation) | IMMA (MG) · Social Navigation | graph accuracy (%): 77.98 -> 81.38 | Improved |
| add explicit disentanglement loss instead of PLT (ablation) | IMMA (MG) · Social Navigation | ADE: 0.148 -> 0.197 | Hurt |
| use multiple latent graph layers instead of a single attentional layer (ablation) | IMMA · Social Navigation | ADE: 0.164 -> 0.148 | Improved |
| add PLT to multiplex graph on PHASE relational inference | IMMA (MG) · PHASE | graph accuracy (%): 79.02 -> 79.21 | No clear effect |
| IMMA MG vs fNRI on NBA | IMMA (MG) · NBA | ADE: 0.804 -> 0.833 | Hurt |
| IMMA (MG+PLT) vs best baselines on trajectory prediction | IMMA · NBA | ADE: 0.804 -> 0.769 | Improved |
| IMMA (MG+PLT) vs RFM on relational inference | IMMA · Social Navigation | graph accuracy (%): 71.53 -> 81.38 | Improved |
| zero-shot test with 2x more agents | IMMA (MG+PLT) · Social Navigation | ADE: 0.310 -> 0.195 | Improved |
Closest papers by meaning
- Generative Attention Networks for Multi-Agent Behavioral Modeling (AAAI 2020 · similarity 0.91)
- VAIN: Attentional Multi-agent Predictive Modeling (NeurIPS 2017 · similarity 0.90)
- Inverse Attention Agents for Multi-Agent Systems (ICLR 2025 · similarity 0.88)
- A Theoretical Study of (Hyper) Self-Attention through the Lens of Interactions: Representation, Training, Generalization (ICML 2025 · similarity 0.87)
- G-Memory: Tracing Hierarchical Memory for Multi-Agent Systems (NeurIPS 2025 · similarity 0.86)
- Mental-Models for Multi-Agent Systems (NeurIPS 2026 · similarity 0.85)
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.
