StreamPhy: Streaming Inference of High-Dimensional Physical Dynamics via State Space Models

TL;DR

StreamPhy is a novel streaming inference framework that accurately and efficiently models high-dimensional physical dynamics from irregular sparse measurements in real time.

Contribution

It introduces a data-adaptive encoder, a structured state-space model, and an expressive FT-FiLM decoder, advancing real-time physical field inference from sparse data.

Findings

Outperforms state-of-the-art methods with at least 48% accuracy improvement.

Achieves 20-100X faster inference than diffusion-based models.

Supports complex dynamics with a richer function class.

Abstract

Inferring the evolution of high-dimensional and multi-modal (e.g., spatio-temporal) physical fields from irregular sparse measurements in real time is a fundamental challenge in science and engineering. Existing approaches, including diffusion-based generative models and functional tensor methods, typically operate in offline settings, depend on full temporal observations, or incur substantial inference cost. We propose StreamPhy, an end-to-end framework that enables efficient and accurate streaming inference of full-field physical dynamics from incoming irregular sparse measurements. The framework integrates a data-adaptive observation encoder that is robust to arbitrary observation patterns, a structured state-space model that supports memory-efficient online updates across irregular time intervals, and an expressive Functional Tensor Feature-wise Linear Modulation (FT-FiLM) decoder for continuous-field generation. We prove that FT-FiLM is more expressive than the functional Tucker model, admitting a richer function class for handling complex dynamics. Experiments on three representative physical systems under challenging sampling patterns show that StreamPhy consistently outperforms state-of-the-art baselines, with at least 48\% improvement in accuracy and up to 20--100X faster inference than diffusion-based methods.