Research map / Time series / Foundation models for time series
DeepDPM: Dynamic Population Mapping via Deep Neural Network
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 |
|---|---|---|---|
| Use stacked SRCNNs instead of Random Forest | Static population mapping · Shanghai weekdays, X1-X2 | RMSE: 47.0408 -> 44.5574 | Helped |
| Use stacked SRCNNs instead of Random Forest | Static population mapping · Shanghai weekdays, X1-X3 | RMSE: 93.5023 -> 78.9081 | Helped |
| Add time embedding to LSTM (ablation) | Dynamic population mapping · Shanghai, X2-X3 | NRMSE: 0.4236 -> 0.3862; reduces NRMSE by $14.5\%$ | Helped |
| Add LSTM temporal modeling to static mapping (ablation) | Dynamic population mapping · Shanghai, X2-X3 | NRMSE: 0.4517 -> 0.4236 | Helped |
| Add time embedding to LSTM (ablation) | Dynamic population mapping · Shanghai, X2-X3 | MAE: 31.60 -> 32.01 | Hurt |
| Segment training data into the 7:00-17:00 period (ablation) | Static Model · Shanghai, X1-X3 | RMSE: 78.9081 -> 76.9615 | Helped |
| Add PoI information to the spatial model (ablation) | DeepDPM static mapping · Shanghai population mapping | Predictive performance | Helped |
| Use entertainment PoIs, particularly with residence PoIs (ablation) | DeepDPM static mapping · Shanghai population mapping | Predictive performance | Helped |
| Train independent low-upscaling SRCNN units and stack them rather than train a cascading end-to-end model (ablation) | DeepDPM spatial mapping · Shanghai population mapping | Prediction accuracy | Helped |
Closest papers by meaning
- Neural Collective Graphical Models for Estimating Spatio-Temporal Population Flow from Aggregated Data (AAAI 2019 · similarity 0.84)
- Leveraging Sparse Observations to Predict Species Abundance Across Space and Time (AAAI 2026 · similarity 0.82)
- Convolutional State Space Models for Long-Range Spatiotemporal Modeling (NeurIPS 2023 · similarity 0.82)
- Population Transformer: Learning Population-level Representations of Neural Activity (ICLR 2025 · similarity 0.81)
- Time Series Generation Under Data Scarcity: A Unified Generative Modeling Approach (NeurIPS 2025 · similarity 0.80)
- POCO: Scalable Neural Forecasting through Population Conditioning (NeurIPS 2025 · similarity 0.80)
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.