A Nonautonomous Equation Discovery Method for Time Signal Classification

DOI10.1137/21M1405216MaRDI QIDQ5024512zbMATH OpenFDO

Authors Ryeongkyung Yoon, Braxton Osting, Harish S. Bhat

Publication date 26 January 2022

Published in SIAM Journal on Applied Dynamical Systems (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2011.11096

zbMATH Keywords

classificationneural networksadjoint methodequation discoverytime signal analysis

Mathematics Subject Classification ID

Sequential statistical analysis (62L10) Artificial neural networks and deep learning (68T07) Approximation methods and numerical treatment of dynamical systems (37M99)

Abstract: Certain neural network architectures, in the infinite-layer limit, lead to systems of nonlinear differential equations. Motivated by this idea, we develop a framework for analyzing time signals based on non-autonomous dynamical equations. We view the time signal as a forcing function for a dynamical system that governs a time-evolving hidden variable. As in equation discovery, the dynamical system is represented using a dictionary of functions and the coefficients are learned from data. This framework is applied to the time signal classification problem. We show how gradients can be efficiently computed using the adjoint method, and we apply methods from dynamical systems to establish stability of the classifier. Through a variety of experiments, on both synthetic and real datasets, we show that the proposed method uses orders of magnitude fewer parameters than competing methods, while achieving comparable accuracy. We created the synthetic datasets using dynamical systems of increasing complexity; though the ground truth vector fields are often polynomials, we find consistently that a Fourier dictionary yields the best results. We also demonstrate how the proposed method yields graphical interpretability in the form of phase portraits.

Recommendations

• A method for nonlinear system classification in the time-frequency plane
• Signal extraction from nonstationary time series
• Nonstationary signal pattern recognition using fast time-time filtering and decision tree
• A nonlinear dynamics method for signal identification
• Signal extraction for finite nonstationary time series
• A new time series classification approach
• scientific article; zbMATH DE number 1163450
• Signal extraction for nonstationary time series with diverse sampling rules

Cites work

• scientific article; zbMATH DE number 51724 (Why is no real title available?)
• Blowup from randomly switching between stable boundary conditions for the heat equation
• Deep learning
• Deep learning for time series classification: a review
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Learning representations by back-propagating errors
• Neural networks and deep learning. A textbook
• Neurons with graded response have collective computational properties like those of two-state neurons
• On the computational power of neural nets
• On the convergence of the SINDy algorithm
• Ordinary differential equations and dynamical systems
• Stable architectures for deep neural networks
• Turing computability with neural nets

Cited in

(3)

• Data-driven learning of nonautonomous systems
• A dynamical system-based framework for dimension reduction
• Deep learning for nonlinear time series: examples for inferring slow driving forces

Describes a project that uses

Uses Software

• torchcde
• NeuralCDE
• GRU-ODE-Bayes
• torchdiffeq
• AntisymmetricRNN

This page was built for publication: A Nonautonomous Equation Discovery Method for Time Signal Classification

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5024512)