Linear RNNs provably learn linear dynamical systems

DOI10.1007/S10957-024-02521-3MaRDI QIDQ6636796zbMATH OpenFDO

Authors Lifu Wang, Tianyu Wang, Shengwei Yi, Bo Shen, Bo Hu, Xing Cao

Publication date 12 November 2024

Published in Journal of Optimization Theory and Applications (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/S10957-024-02521-3

zbMATH Keywords

linear systemsmachine learningnon-convex optimizationrecurrent neural networks

Mathematics Subject Classification ID

Calculus of variations and optimal control; optimization (49-XX) Mathematical programming (90Cxx)

Recommendations

• Metric entropy limits on recurrent neural network learning of linear dynamical systems
• Modeling of continuous time dynamical systems with input by recurrent neural networks
• Distributed sequence memory of multidimensional inputs in recurrent networks
• Understanding recurrent neural networks using nonequilibrium response theory
• Gradient descent learns linear dynamical systems

Cites work

• Finite sample properties of system identification methods
• On the computational power of neural nets

This page was built for publication: Linear RNNs provably learn linear dynamical systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6636796)