Counting equilibria in a random non-gradient dynamics with heterogeneous relaxation rates
Bertrand Lacroix-A-Chez-Toine, Yan V Fyodorov

TL;DR
This paper analyzes the number and stability of equilibria in a complex nonlinear random system with heterogeneous relaxation rates, revealing phase transitions and universal behaviors using advanced mathematical techniques.
Contribution
It extends previous models by deriving exact analytical results for systems with a spectrum of relaxation rates, revealing new phase transitions and universal properties.
Findings
Identifies a topology trivialisation transition from many to a single equilibrium.
Shows a phase with exponentially many stable equilibria depending on the gradient component.
Finds the complexity behavior at transitions depends only on small relaxation rate spectrum details.
Abstract
We consider a nonlinear autonomous random dynamical system of degrees of freedom coupled by Gaussian random interactions and characterized by a continuous spectrum of real positive relaxation rates. Using Kac-Rice formalism, the computation of annealed complexities (both of stable equilibria and of all types of equilibria) is reduced to evaluating the averages involving the modulus of the determinant of the random Jacobian matrix. In the limit of large system we derive exact analytical results for the complexities for short-range correlated coupling fields, extending results previously obtained for the "homogeneous" relaxation spectrum characterised by a single relaxation rate. We show the emergence of a "topology trivialisation" transition from a complex phase with exponentially many equilibria to a simple phase with a single equilibrium as the magnitude…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
