Research map / Time series / Causal and generative time series
RNN with Particle Flow for Probabilistic Spatio-temporal Forecasting
Soumyasundar Pal, Liheng Ma, Yingxue Zhang, Mark Coates
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 |
|---|---|---|---|
| Use graph-aware AGCGRU dynamics with particle flow for probabilistic forecasting | AGCGRU+flow · PeMSD3 | MAE: 14.42/15.87/17.10/18.29 -> 13.79/14.84/15.58/16.06 | Improved |
| Use graph-aware AGCGRU dynamics with particle flow for probabilistic forecasting | AGCGRU+flow · PeMSD4 | CRPS: DeepAR: 1.13/1.52/1.82/2.07; AGCGRU+flow: 1.08/1.32/1.46/1.56 | Improved |
| Use GRU dynamics with particle flow instead of deterministic FC-GRU | GRU+flow · Electricity | Normalized Deviation: FC-GRU: 0.102/0.118/0.098/0.203; GRU+flow: 0.070/0.071/0.069/0.140 | Improved |
| Use GRU dynamics with particle flow for probabilistic forecasting | GRU+flow · Electricity, Traffic, Taxi, and Wikipedia | CRPS_sum: Electricity 0.013; Traffic 0.028; Taxi 0.140; Wikipedia 0.054 | Improved |
| Replace particle flow with a Bootstrap Particle Filter (ablation) | AGCGRU+flow · PeMSD3 | MAE: 13.79/14.84/15.58/16.06 -> 14.19/15.13/15.85/16.35 | Hurt |
| Replace particle flow with a Bootstrap Particle Filter (ablation) | AGCGRU+flow · PeMSD3 | CRPS: 10.53/11.39/12.03/12.47 -> 11.32/11.94/12.55/12.92 | Hurt |
| Replace probabilistic flow forecasting with a deterministic encoder-decoder (ablation) | AGCGRU+flow · PeMSD3 | MAE: 13.79/14.84/15.58/16.06 -> 13.96/15.37/16.52/17.45 | Hurt |
| Use DCGRU rather than AGCGRU for state transitions (ablation) | AGCGRU+flow · PeMSD3 | MAE: 13.79/14.84/15.58/16.06 -> 14.48/15.67/16.52/17.36 | Hurt |
| Use GRU rather than AGCGRU for state transitions (ablation) | AGCGRU+flow · PeMSD3 | MAE: 13.79/14.84/15.58/16.06 -> 14.40/16.10/17.63/19.18 | Hurt |
| Use GRU dynamics with particle flow rather than FC-GRU | GRU+flow · Traffic | Normalized Deviation: FC-GRU: 0.259/0.528/0.233; GRU+flow: 0.133/0.322/0.125 | Mixed |
Closest papers by meaning
- Probabilistic Forecasting with Stochastic Interpolants and Föllmer Processes (ICML 2024 · similarity 0.86)
- Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting (ICLR 2025 · similarity 0.86)
- ProbTS: Benchmarking Point and Distributional Forecasting across Diverse Prediction Horizons (NeurIPS 2024 · similarity 0.86)
- ST-FiT: Inductive Spatial-Temporal Forecasting with Limited Training Data (AAAI 2025 · similarity 0.85)
- How Different from the Past? Spatio-Temporal Time Series Forecasting with Self-Supervised Deviation Learning (NeurIPS 2025 · similarity 0.85)
- Expand and Compress: Exploring Tuning Principles for Continual Spatio-Temporal Graph Forecasting (ICLR 2025 · similarity 0.85)
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.
