Overview

Dynamical systems describe how a system's state changes over time and explain behaviors such as equilibria, oscillations, attractors, bifurcations, and chaos. Researchers analyze these behaviors mathematically and learn models from data using approaches such as Koopman operators, neural ODEs, physics-informed networks, and structure-preserving Hamiltonian models. Applications range from forecasting Lorenz and pendulum dynamics to modeling reaction-diffusion processes, neural activity, and robot motion, often with control and safety guarantees. A central challenge is that accurate short-term predictions do not ensure faithful long-term behavior: learned models can drift, violate physical constraints, or fail to generalize to unseen regimes.

Latest papers

515 papers34 papers in the last four weeks, down 43% on the four weeks before. 0.3% of all new papers.

Oct 14, 2024

Automated Discovery of Operable Dynamics from Videos

Dynamical systems form the foundation of scientific discovery, traditionally modeled with predefined state variables such as the angle and angular velocity, and differential equations such as the equation of motion for a single pendulum. We introduce a framework that automatically discovers a low-dimensional and operable representation of system dynamics, including a set of compact state variables that preserve the smoothness of the system dynamics and a differentiable vector field, directly from video without requiring prior domain-specific knowledge. The prominence and effectiveness of the proposed approach are demonstrated through both quantitative and qualitative analyses of a range of dynamical systems, including the identification of stable equilibria, the prediction of natural frequencies, and the detection of chaotic and limit cycle behaviors. The results highlight the potential of our data-driven approach to advance automated scientific discovery.
Aug 30, 2024

Learning Maximal Regions of Attraction on Compact Sets using Taylor Lyapunov Networks

We consider the problem of learning maximal Lyapunov functions to compute the maximal region of attraction of nonlinear dynamical systems using neural networks. Two obstacles limit existing neural-network approaches: 1) finite-width approximations can fail to preserve the strict Lyapunov inequalities near the equilibrium, where they become arbitrarily weak, and 2) training is necessarily performed on a bounded computational domain for which the classical Lyapunov and Zubov theorems do not directly apply. To address these issues, first we derive Lyapunov and Zubov theorems directly on a prescribed compact computational domain, characterizing the maximal region of attraction within that domain as the solution of a boundary-value PDE. Second, we introduce the Taylor Lyapunov network, an architecture that prescribes the quadratic leading-order term via the system's linearization and represents the higher-order remainder through a structured cubic factorization, guaranteeing that any network in this class is a valid local Lyapunov function by construction. Finally, we prove that Taylor Lyapunov networks are universal approximators of maximal Lyapunov functions on compact domains, and combine these results into an unsupervised, physics-informed training procedure. Numerical experiments on systems of varying dimensions and stability types show that the proposed architecture is efficient at estimating the maximum region of attraction.
Jun 7, 2024

Gradient Descent on Logistic Regression with Non-Separable Data and Large Step Sizes

We study gradient descent (GD) dynamics on logistic regression problems with large, constant step sizes. For linearly-separable data, it is known that GD converges to the minimizer with arbitrarily large step sizes, a property which no longer holds when the problem is not separable. In fact, the behaviour can be much more complex -- a sequence of period-doubling bifurcations begins at the critical step size 2/λ2/λ, where λλ is the largest eigenvalue of the Hessian at the solution. Using a smaller-than-critical step size guarantees convergence if initialized nearby the solution: but does this suffice globally? In one dimension, we show that a step size less than 1/λ1/λ suffices for global convergence. However, for all step sizes between 1/λ1/λ and the critical step size 2/λ2/λ, one can construct a dataset such that GD converges to a stable cycle. In higher dimensions, this is actually possible even for step sizes less than 1/λ1/λ. Our results show that although local convergence is guaranteed for all step sizes less than the critical step size, global convergence is not, and GD may instead converge to a cycle depending on the initialization.
Oct 28, 2021

Roto-translated Local Coordinate Frames For Interacting Dynamical Systems

Modelling interactions is critical in learning complex dynamical systems, namely systems of interacting objects with highly non-linear and time-dependent behaviour. A large class of such systems can be formalized as geometric graphs\textit{geometric graphs}, i.e.\textit{i.e.}, graphs with nodes positioned in the Euclidean space given an arbitrarily\textit{arbitrarily} chosen global coordinate system, for instance vehicles in a traffic scene. Notwithstanding the arbitrary global coordinate system, the governing dynamics of the respective dynamical systems are invariant to rotations and translations, also known as Galilean invariance\textit{Galilean invariance}. As ignoring these invariances leads to worse generalization, in this work we propose local coordinate frames per node-object to induce roto-translation invariance to the geometric graph of the interacting dynamical system. Further, the local coordinate frames allow for a natural definition of anisotropic filtering in graph neural networks. Experiments in traffic scenes, 3D motion capture, and colliding particles demonstrate that the proposed approach comfortably outperforms the recent state-of-the-art.
May 22, 2021

Universal set of Observables for Forecasting Physical Systems through Causal Embedding

We show how a pair of points can uniquely represent a left-infinite sequence obtained from observations of an underlying dynamical system through a phenomenon called causal embedding. A driven dynamical system creates such pairs, and a function can be learnt on them that can reconstruct the underlying dynamics as in Takens delay embedding. The approach assures embedding stability unlike Takens delay embedding, and learnability which can be absent while stability can be present in the current reservoir computing framework. This accurately models underlying systems where recent methods like the next-generation reservoir computing fail. We demonstrate results and compare with the other methods including SINDY-PI.