Research map / Probabilistic and scientific modeling / Dynamical systems
Discovering Intrinsic Spatial-Temporal Logic Rules to Explain Human Actions
Chengzhi Cao, Chao Yang, Ruimao Zhang, Shuang 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 |
|---|---|---|---|
| compare against Y-Net on SDD | Ours · SDD | ADE20: 7.85 -> 6.41 | Improved |
| compare against Y-Net on SDD | Ours · SDD | FDE20: 11.85 -> 10.23 | Improved |
| compare against MID on SDD | Ours · SDD | ADE20: 7.61 -> 6.41 | Improved |
| compare against MID on SDD | Ours · SDD | FDE20: 14.30 -> 10.23 | Improved |
| compare against NSP-SFM on SDD (F1) | Ours · SDD | F1 score: 0.58 -> 0.59 | Improved |
| compare against Y-Net on NBA 1.0s | Ours · NBA SportVU | ADE20: 0.38 -> 0.30 | Improved |
| compare against Y-Net on NBA event accuracy | Ours · NBA SportVU | accuracy: 0.69 -> 0.73 | Improved |
| compare against NSP-SFM on NBA vs SDD | NSP-SFM · NBA SportVU | ADE20: 0.41 -> 0.30 at 1.0s | No clear effect |
| add noise and randomly remove tracks in NBA (ablation) | Ours · NBA SportVU 1.0s | ADE: 0.30 -> 0.31 | Hurt |
| add noise and randomly remove tracks in NBA (ablation) | Ours · NBA SportVU 4.0s | ADE: 1.13 -> 1.18 | Hurt |
| add noise and randomly remove tracks in NBA (ablation) | Y-Net · NBA SportVU 1.0s | ADE: 0.38 -> 0.51 | Hurt |
| learn rule weights on 4 synthetic datasets with known rules | rule weight learning · Synthetic TLPP datasets | MAE of rule weights | Improved |
Closest papers by meaning
- Learning Human Habits with Rule-Guided Active Inference (ICLR 2026 · similarity 0.87)
- Evolving Minds: Logic-Informed Inference from Temporal Action Patterns (ICML 2025 · similarity 0.87)
- Explaining Point Processes by Learning Interpretable Temporal Logic Rules (ICLR 2022 · similarity 0.86)
- Sports-Traj: A Unified Trajectory Generation Model for Multi-Agent Movement in Sports (ICLR 2025 · similarity 0.86)
- PSI: A Benchmark for Human Interpretation and Response in Traffic Interactions (NeurIPS 2025 · similarity 0.85)
- RHYTHM: Reasoning with Hierarchical Temporal Tokenization for Human Mobility (NeurIPS 2025 · 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.
