PySHRED: A Python package for SHallow REcurrent Decoding for sparse sensing, model reduction and scientific discovery

2025-07-28

AI Summary

This work addresses the challenge of modeling high-dimensional, multiscale, nonlinear, and parametric dynamical systems from sparse spatiotemporal observational data. We propose SHRED—a modular, scalable, shallow recursive decoding framework—that unifies sparse-aware representation learning, nonlinear dimensionality reduction, and physics-informed modeling. By synergistically integrating lightweight recursive architectures with adaptive data preprocessing modules, SHRED significantly enhances robustness to noise and fidelity in capturing multiscale dynamic couplings. Its core innovation lies in embedding scientific discovery—such as conservation law identification and governing equation learning—directly into the reduced-order modeling pipeline, thereby enabling interpretable and generalizable, data-driven dynamical system learning. We release an open-source MIT-licensed toolkit featuring comprehensive documentation and cross-domain validation on fluid dynamics and reaction–diffusion systems, supporting efficient reproducibility and plug-and-play scientific discovery.

Technology Category

• Machine Learning: Multimodal Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

SHallow REcurrent Decoders (SHRED) provide a deep learning strategy for modeling high-dimensional dynamical systems and/or spatiotemporal data from dynamical system snapshot observations. PySHRED is a Python package that implements SHRED and several of its major extensions, including for robust sensing, reduced order modeling and physics discovery. In this paper, we introduce the version 1.0 release of PySHRED, which includes data preprocessors and a number of cutting-edge SHRED methods specifically designed to handle real-world data that may be noisy, multi-scale, parameterized, prohibitively high-dimensional, and strongly nonlinear. The package is easy to install, thoroughly-documented, supplemented with extensive code examples, and modularly-structured to support future additions. The entire codebase is released under the MIT license and is available at https://github.com/pyshred-dev/pyshred.

Problem

• Modeling high-dimensional dynamical systems from sparse data
• Handling noisy, multi-scale, and nonlinear real-world data
• Providing tools for robust sensing and physics discovery

Innovation

• Python package for SHallow REcurrent Decoding
• Deep learning for high-dimensional dynamical systems
• Handles noisy, multi-scale, nonlinear real-world data