DeepAI AI Chat
Log In Sign Up

Decomposed Linear Dynamical Systems (dLDS) for learning the latent components of neural dynamics

by   Noga Mudrik, et al.
Georgia Institute of Technology
Johns Hopkins University

Learning interpretable representations of neural dynamics at a population level is a crucial first step to understanding how neural activity relates to perception and behavior. Models of neural dynamics often focus on either low-dimensional projections of neural activity, or on learning dynamical systems that explicitly relate to the neural state over time. We discuss how these two approaches are interrelated by considering dynamical systems as representative of flows on a low-dimensional manifold. Building on this concept, we propose a new decomposed dynamical system model that represents complex non-stationary and nonlinear dynamics of time-series data as a sparse combination of simpler, more interpretable components. The decomposed nature of the dynamics generalizes over previous switched approaches and enables modeling of overlapping and non-stationary drifts in the dynamics. We further present a dictionary learning-driven approach to model fitting, where we leverage recent results in tracking sparse vectors over time. We demonstrate that our model can learn efficient representations and smooth transitions between dynamical modes in both continuous-time and discrete-time examples. We show results on low-dimensional linear and nonlinear attractors to demonstrate that our decomposed dynamical systems model can well approximate nonlinear dynamics. Additionally, we apply our model to C. elegans data, illustrating a diversity of dynamics that is obscured when classified into discrete states.


page 20

page 21


Learning interpretable continuous-time models of latent stochastic dynamical systems

We develop an approach to learn an interpretable semi-parametric model o...

Variational approach for learning Markov processes from time series data

Inference, prediction and control of complex dynamical systems from time...

Charts and atlases for nonlinear data-driven models of dynamics on manifolds

We introduce a method for learning minimal-dimensional dynamical models ...

Interpretable statistical representations of neural population dynamics and geometry

The dynamics of neuron populations during diverse tasks often evolve on ...

Extraction of Discrete Spectra Modes from Video Data Using a Deep Convolutional Koopman Network

Recent deep learning extensions in Koopman theory have enabled compact, ...

Robust and Scalable Models of Microbiome Dynamics

Microbes are everywhere, including in and on our bodies, and have been s...

Code Repositories


Decomposed Linear Dynamical Systems (dLDS) for learning the latent components of neural dynamics

view repo