Research map / Multi-agent systems and games / Human-AI and goal-conditioned
VectorNet: Encoding HD Maps and Agent Dynamics From Vectorized Representation
Check 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 |
|---|---|---|---|
| represent map and agent histories as vectors and encode them with a hierarchical graph instead of rasterized images and a ConvNet | behavior prediction · Argoverse validation | DE@3s: 4.49 -> 3.67 | Helped |
| represent map and agent histories as vectors and encode them with a hierarchical graph instead of rasterized images and a ConvNet | behavior prediction · In-house validation | DE@3s: 1.00 -> 1.00 | No clear effect |
| add HD map context to target-agent history (ablation) | VectorNet · In-house validation | DE@3s: 1.29 -> 1.11 | Helped |
| add HD map context to target-agent history (ablation) | VectorNet · Argoverse validation | DE@3s: 5.24 -> 3.94 | Helped |
| add other-agent trajectory vectors to map context (ablation) | VectorNet · In-house validation | DE@3s: 1.11 -> 1.05 | Helped |
| add other-agent trajectory vectors to map context (ablation) | VectorNet · Argoverse validation | DE@3s: 3.94 -> 3.84 | Helped |
| add graph completion auxiliary objective (ablation) | VectorNet with map and agent context · In-house validation | DE@3s: 1.05 -> 1.00 | Helped |
| add graph completion auxiliary objective (ablation) | VectorNet with map and agent context · Argoverse validation | DE@3s: 3.84 -> 3.67 | Helped |
| increase polyline subgraph depth from 1 to 3 (ablation) | VectorNet · In-house validation | DE@3s: 1.09 -> 1.00 | Helped |
| increase polyline subgraph depth from 1 to 3 (ablation) | VectorNet · Argoverse validation | DE@3s: 3.89 -> 3.67 | Helped |
| increase polyline subgraph width from 64 to 128 (ablation) | VectorNet · Argoverse validation | DE@3s: 3.67 -> 3.93 | Hurt |
| increase global graph depth from 1 to 2 (ablation) | VectorNet · Argoverse validation | DE@3s: 3.67 -> 3.69 | No clear effect |
Closest papers by meaning
- HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention (CVPR 2024 · similarity 0.88)
- Density-Adaptive Model Based on Motif Matrix for Multi-Agent Trajectory Prediction (CVPR 2024 · similarity 0.87)
- SEPT: Towards Efficient Scene Representation Learning for Motion Prediction (ICLR 2024 · similarity 0.87)
- Graph-Based Prediction and Planning Policy Network (GP3Net) for Scalable Self-Driving in Dynamic Environments Using Deep Reinforcement Learning (AAAI 2024 · similarity 0.87)
- Scenario Dreamer: Vectorized Latent Diffusion for Generating Driving Simulation Environments (CVPR 2025 · similarity 0.87)
- DriveGPT: Scaling Autoregressive Behavior Models for Driving (ICML 2025 · 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.