Differential equations for continuous-time deep learning

DOI10.1090/NOTI2930MaRDI QIDQ6610701zbMATH OpenFDO

Authors Lars Ruthotto

Publication date 26 September 2024

Published in Notices of the American Mathematical Society (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1090/NOTI2930

Mathematics Subject Classification ID

PDE constrained optimization (numerical aspects) (49M41) Mean field games and control (49N80) Applications of optimal control and differential games (49N90) Numerical optimization and variational techniques (65K10) Artificial neural networks and deep learning (68T07)

Recommendations

• A proposal on machine learning via dynamical systems
• Deep learning via dynamical systems: an approximation perspective
• A functional approach to interpreting the role of the adjoint equation in machine learning
• The unreasonable effectiveness of deep learning in artificial intelligence
• An introduction to the mathematics of deep learning

Cited in

(3)

• A functional approach to interpreting the role of the adjoint equation in machine learning
• Vanilla feedforward neural networks as a discretization of dynamical systems
• Deep networks for system identification: a survey

This page was built for publication: Differential equations for continuous-time deep learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6610701)