Cross-Domain Long-Term Forecasting: Radiation Dose from Sparse Neutron Sensor via Spatio-Temporal Operator Network

2025-10-20

AI Summary

Predicting long-term physical fields from sparse, cross-domain, and multi-scale sensor data remains challenging—particularly when sensing and prediction occur on distinct physical manifolds and over extended temporal horizons; existing methods are constrained by requirements for dense collocated data and short-term contextual windows. Method: We propose STONe, a non-autoregressive spatiotemporal operator network, enabling direct end-to-end nonlinear mapping from sparse ground-level neutron monitoring data to high-altitude radiation dose fields across manifolds. STONe overcomes classical operator learning limitations—namely, domain alignment between input/output or iterative autoregressive inference. Contribution/Results: Trained on 23 years of global neutron data, STONe achieves 180-day lead-time forecasting with high accuracy and millisecond-level inference latency. It establishes a novel paradigm for real-time, physics-informed modeling of complex spatiotemporal fields in geophysical and climate systems.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Forecasting unobservable physical quantities from sparse, cross-domain sensor data is a central unsolved problem in scientific machine learning. Existing neural operators and large-scale forecasters rely on dense, co-located input-output fields and short temporal contexts, assumptions that fail in real-world systems where sensing and prediction occur on distinct physical manifolds and over long timescales. We introduce the Spatio-Temporal Operator Network (STONe), a non-autoregressive neural operator that learns a stable functional mapping between heterogeneous domains. By directly inferring high-altitude radiation dose fields from sparse ground-based neutron measurements, STONe demonstrates that operator learning can generalize beyond shared-domain settings. It defines a nonlinear operator between sensor and target manifolds that remains stable over long forecasting horizons without iterative recurrence. This challenges the conventional view that operator learning requires domain alignment or autoregressive propagation. Trained on 23 years of global neutron data, STONe achieves accurate 180-day forecasts with millisecond inference latency. The framework establishes a general principle for cross-domain operator inference, enabling real-time prediction of complex spatiotemporal fields in physics, climate, and energy systems.

Problem

• Forecasting unobservable radiation dose from sparse neutron sensor data
• Learning stable mappings between heterogeneous physical domains
• Enabling long-term predictions without domain alignment requirements

Innovation

• Non-autoregressive neural operator for cross-domain mapping
• Direct inference from sparse sensors to target fields
• Stable long-term forecasting without domain alignment