Research map / Multi-agent systems and games / Human-AI and goal-conditioned
SoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints
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 |
|---|---|---|---|
| add both social and physical attention to the generative model (ablation) | SoPhie · ETH-UCY AVG | ADE: 0.58 -> 0.54 | Helped |
| add both social and physical attention to the generative model (ablation) | SoPhie · SDD | ADE: 19.25 -> 16.27 | Helped |
| add physical attention to image features (ablation) | SoPhie · ETH-UCY AVG | ADE: 0.61 -> 0.57 | Helped |
| add social attention to trajectory features (ablation) | SoPhie · ETH-UCY AVG | ADE: 0.61 -> 0.58 | Helped |
| add either physical attention or social attention rather than neither (ablation) | SoPhie · SDD | ADE: 18.40 -> 16.52; 18.40 -> 17.57 | Helped |
| add both attention mechanisms rather than physical attention only (ablation) | SoPhie · SDD | ADE: 16.52 -> 16.27 | Helped |
| add physical attention to the trajectory-only model (ablation) | SoPhie · ETH-UCY AVG | ADE: 0.61 -> 0.54 | Helped |
| add physical attention to the trajectory-only model (ablation) | SoPhie · SDD | ADE: 17.76 -> 16.27 | Helped |
| use social and physical context without attention (ablation) | SoPhie · ETH-UCY AVG | ADE: 0.54 -> 0.61 | Hurt |
| use social and physical context without attention (ablation) | SoPhie · SDD | ADE: 16.27 -> 18.40 | Hurt |
| use the complete model rather than social features and social attention only (ablation) | SoPhie · ETH-UCY AVG | ADE: 0.61 -> 0.54 | Helped |
| use the complete model rather than social features and social attention only (ablation) | SoPhie · SDD | ADE: 17.76 -> 16.27 | Helped |
Closest papers by meaning
- Social-BiGAT: Multimodal Trajectory Forecasting using Bicycle-GAN and Graph Attention Networks (NeurIPS 2019 · similarity 0.91)
- Social GAN: Socially Acceptable Trajectories With Generative Adversarial Networks (CVPR 2018 · similarity 0.91)
- Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction (CVPR 2020 · similarity 0.88)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.86)
- Higher-order Relational Reasoning for Pedestrian Trajectory Prediction (CVPR 2024 · similarity 0.85)
- SocialCircle: Learning the Angle-based Social Interaction Representation for Pedestrian Trajectory Prediction (CVPR 2024 · similarity 0.84)
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.