Predicting the Dynamics of Complex System via Multiscale Diffusion Autoencoder

Abstract

Predicting the dynamics of complex systems is crucial for various scientific and engineering applications. The accuracy of predictions depends on the model's ability to capture the intrinsic dynamics. While existing methods capture key dynamics by encoding a low-dimensional latent space, they overlook the inherent multiscale structure of complex systems, making it difficult to accurately predict complex spatiotemporal evolution. Therefore, we propose a Multiscale Diffusion Prediction Network (MDPNet) that leverages the multiscale structure of complex systems to discover the latent space of intrinsic dynamics. First, we encode multiscale features through a multiscale diffusion autoencoder to guide the diffusion model for reliable reconstruction. Then, we introduce an attention-based graph neural ordinary differential equation to model the co-evolution across different scales. Extensive evaluations on representative systems demonstrate that the proposed method achieves an average prediction error reduction of 53.23% compared to baselines, while also exhibiting superior robustness and generalization.

KDD 2025

References in corpus (7)

• Diffusion Models in Vision: A Survey
• Learning data driven discretizations for partial differential equations
• A Physics-informed Diffusion Model for High-fidelity Flow Field Reconstruction
• Mutual Information, Neural Networks and the Renormalization Group
• Multiscale unfolding of real networks by geometric renormalization
• The low-rank hypothesis of complex systems
• PDEBENCH: An Extensive Benchmark for Scientific Machine Learning