Research map / Graph learning / Knowledge and temporal graphs
GeoFlow: Geo-Aware Modeling of Inter-Area Relationships in Origin-Destination Flow Prediction and Generation
Zherui Huang, Guanjie Zheng, Hao Xue, Linghe Kong
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 |
|---|---|---|---|
| add relative coordinates, k-hop and free-flow distance features to straight-line distance (ablation) | GeoFlow generation · Commuting OD flow generation | CPC: 0.462 -> 0.527 | Improved |
| add relative coordinates, k-hop and free-flow distance features to straight-line distance (ablation) | GeoFlow generation · Commuting OD flow generation | NRMSE/JSD: 6.162/0.252 -> 4.330/0.194 | Improved |
| add geospatial features to straight-line distance (ablation) | GeoFlow prediction · Commuting OD flow prediction | CPC/NRMSE: CPC 0.594 -> 0.597; NRMSE 1.994 -> 1.730 | Mixed |
| add the proposed geo-conditioned encoder and axial-global attention decoder (ablation) | GeoFlow prediction · Commuting OD flow prediction | CPC/NRMSE/JSD: 0.567/2.134/0.201 -> 0.597/1.730/0.171 | Improved |
| add the proposed geo-conditioned encoder and axial-global attention decoder (ablation) | GeoFlow generation · Commuting OD flow generation | CPC/NRMSE/JSD: 0.510/6.780/0.209 -> 0.527/4.330/0.194 | Improved |
| add axial-global attention decoder to encoder-only model (ablation) | GeoFlow generation · Commuting OD flow generation | NRMSE: 6.625 -> 4.330 | Improved |
| add geo-conditioned encoder to decoder-only model (ablation) | GeoFlow prediction · Commuting OD flow prediction | NRMSE: 2.092 -> 1.730 | Improved |
| use GeoFlow prediction instead of TransFlower | OD flow prediction · Commuting | CPC: 0.486 -> 0.604 | Improved |
| use GeoFlow generation instead of WEDAN | OD flow generation · Air Travel | CPC: 0.486 -> 0.457 | Hurt |
| infer relative angle, k-hop and free-flow cues from adjacency and straight-line distance with an MLP | MLP probe · Geographic information reconstruction | relative error (angles): 51.1% average relative error | Hurt |
Closest papers by meaning
- Sat2Flow: A Structure-Aware Diffusion Framework for Human Flow Generation from Satellite Imagery (AAAI 2026 · similarity 0.90)
- Satellites Reveal Mobility: A Commuting Origin-destination Flow Generator for Global Cities (NeurIPS 2025 · similarity 0.88)
- Explainable Origin-Destination Crowd Flow Interpolation via Variational Multi-Modal Recurrent Graph Auto-Encoder (AAAI 2024 · similarity 0.86)
- GeoRouter: Dynamic Paradigm Routing for Worldwide Image Geolocalization (NeurIPS 2026 · similarity 0.85)
- SeoulMMOD: A Large-Scale Multimodal Origin-Destination Flow Benchmark (NeurIPS 2026 · similarity 0.85)
- TorchSpatial: A Location Encoding Framework and Benchmark for Spatial Representation Learning (NeurIPS 2024 · similarity 0.83)
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.
