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

2026-05-08

AI Summary

This work addresses the challenge of real-time inference of high-dimensional, multimodal physical field evolution from irregular and sparse measurements by proposing StreamPhy, an end-to-end streaming inference framework. StreamPhy integrates a data-adaptive observation encoder, a structured state space model, and a novel Functional Tensor Feature-wise Linear Modulation (FT-FiLM) decoder, enabling online updates under arbitrary observation patterns. Theoretical analysis demonstrates that FT-FiLM possesses strictly greater expressive power than functional Tucker models. Experimental results across three distinct physical systems show that StreamPhy improves inference accuracy by at least 48% while achieving 20–100× faster runtime compared to diffusion-based approaches.

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.

Problem

• streaming inference
• high-dimensional physical dynamics
• irregular sparse measurements
• real-time reconstruction
• spatio-temporal fields

Innovation

• streaming inference
• state space models
• functional tensor
• FT-FiLM
• physical dynamics