Research map / Graph learning / Knowledge and temporal graphs
DSTAGNN: Dynamic Spatial-Temporal Aware Graph Neural Network for Traffic Flow Forecasting
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 |
|---|---|---|---|
| Replace pre-defined static adjacency graph with data-driven STAG (STAD-based) (ablation) | DSTAGNN graph structure · PEMS03/04/07/08 | MAE: PEMS03 15.61->15.57; PEMS04 19.41->19.30; PEMS07 21.67->21.42; PEMS08 15.90->15.67 | Helped |
| Full DSTAGNN (STAG + spatial-temporal attention + M-GTU) vs baselines | DSTAGNN · PEMS03/04/07/08 | MAE, MAPE, RMSE: Best MAE on all four PEMS datasets vs best baseline (e.g., PEMS04 19.30 vs 19.50) | Helped |
| Remove spatial-temporal attention mechanism (RemSTA) (ablation) | DSTAGNN · PEMS04 | MAE: qualitative; plotted in Figure 4 only | Hurt |
| Replace multi-head attention with single-head attention (RemM-A) (ablation) | DSTAGNN · PEMS04 | MAE: qualitative; Figure 4 only | Hurt |
| Replace multi-scale GTU with traditional convolution (RemM-GTU) (ablation) | DSTAGNN · PEMS04 | MAE: qualitative; Figure 4 only | Hurt |
| Remove residual connection of ST block outputs (RemRC-OUT) (ablation) | DSTAGNN · PEMS04 | MAE: qualitative; Figure 4 only | Hurt |
Closest papers by meaning
- Spatio-Temporal Pivotal Graph Neural Networks for Traffic Flow Forecasting (AAAI 2024 · similarity 0.91)
- Towards Dynamic Spatial-Temporal Graph Learning: A Decoupled Perspective (AAAI 2024 · similarity 0.91)
- Efficient Traffic Prediction Through Spatio-Temporal Distillation (AAAI 2025 · similarity 0.91)
- TESTAM: A Time-Enhanced Spatio-Temporal Attention Model with Mixture of Experts (ICLR 2024 · similarity 0.90)
- Spatiotemporal-aware Trend-Seasonality Decomposition Network for Traffic Flow Forecasting (AAAI 2025 · similarity 0.89)
- Designing Specialized Two-Dimensional Graph Spectral Filters for Spatial-Temporal Graph Modeling (AAAI 2025 · similarity 0.89)
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.