OmniST: Topology-Invariant Spatiotemporal Forecasting with Guaranteed-Coverage Uncertainty at Unseen Sensor Locations
Abstract
Standard spatiotemporal graph convolutional networks need a fixed-size adjacency matrix, coupling model weights to a specific sensor count N; they cannot produce any output at a sensor absent from training. We present OmniST, a topology-invariant forecaster whose computation has zero structural dependency on the number of nodes, and which, composed with inductive split-conformal calibration, gives distribution-free, coverage-guaranteed prediction intervals at arbitrary (including never-seen) sensor locations, preserving that guarantee as sensors are added or removed. Topology invariance follows from coordinate-based tokenization, k-nearest-neighbor attention from geographic coordinates at each forward pass, and per-node temporal processing via a bidirectional GRU. On real Delhi air-quality data (37 stations, 2019-2023), OmniST attains a 30-day PM2.5 MAE of 31.0 ug/m3 and degrades at most 1.1 percent when 40 percent of stations are removed. Its inductive forecasts at unseen stations beat nearest-neighbor interpolation (about 12 percent) and a learned kriging GNN (IGNNK, about 21 percent), and, unlike either, carry a finite-sample coverage guarantee. The guarantee holds at 90 percent or above under 30-50 percent node dropout across four real datasets over three domains (Delhi and Beijing air quality, NOAA GHCN weather, Caltrans PeMSD7-M traffic), and still holds at 325 real sensors (full PeMS-BAY). We also fix a temporal-drift failure: static conformal under-covers over time (about 0.50 late), but a rolling adaptive-conformal estimator restores target coverage (0.89 overall, 0.96 late) at the cost of wider intervals.
Related
- A Critical Audit of Spatiotemporal Forecasting Benchmark Datasets and Models
- Event-Aware Spatiotemporal Precipitation Forecasting with Geographic Context and Physics-Guided Regularization
- Beyond Point Forecasts: A Survey on Probabilistic Forecasting for Time Series and Spatiotemporal Data
- Explainable spatiotemporal Graph Neural Networks: Multi-City PM2.5 forecasting in Bangladesh with geographic graph topology validation
- Generative diffusion models for spatiotemporal influenza forecasting
BibTeX
@misc{vashishtha2026omnist,
title = {{OmniST: Topology-Invariant Spatiotemporal Forecasting with Guaranteed-Coverage Uncertainty at Unseen Sensor Locations}},
author = {Praveg Vashishtha},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.04256},
eprint = {2611.04256},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.04256},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - OmniST: Topology-Invariant Spatiotemporal Forecasting with Guaranteed-Coverage Uncertainty at Unseen Sensor Locations
AU - Praveg Vashishtha
PY - 2026
DA - 2026/11/03
PB - arXiv
DO - 10.48550/arxiv.2611.04256
UR - https://arxiv.org/abs/2611.04256
AB - Standard spatiotemporal graph convolutional networks need a fixed-size adjacency matrix, coupling model weights to a specific sensor count N; they cannot produce any output at a sensor absent from training. We present OmniST, a topology-invariant forecaster whose computation has zero structural dependency on the number of nodes, and which, composed with inductive split-conformal calibration, gives distribution-free, coverage-guaranteed prediction intervals at arbitrary (including never-seen) sensor locations, preserving that guarantee as sensors are added or removed. Topology invariance follows from coordinate-based tokenization, k-nearest-neighbor attention from geographic coordinates at each forward pass, and per-node temporal processing via a bidirectional GRU. On real Delhi air-quality data (37 stations, 2019-2023), OmniST attains a 30-day PM2.5 MAE of 31.0 ug/m3 and degrades at most 1.1 percent when 40 percent of stations are removed. Its inductive forecasts at unseen stations beat nearest-neighbor interpolation (about 12 percent) and a learned kriging GNN (IGNNK, about 21 percent), and, unlike either, carry a finite-sample coverage guarantee. The guarantee holds at 90 percent or above under 30-50 percent node dropout across four real datasets over three domains (Delhi and Beijing air quality, NOAA GHCN weather, Caltrans PeMSD7-M traffic), and still holds at 325 real sensors (full PeMS-BAY). We also fix a temporal-drift failure: static conformal under-covers over time (about 0.50 late), but a rolling adaptive-conformal estimator restores target coverage (0.89 overall, 0.96 late) at the cost of wider intervals.
ER -