Modulated logarithmic Sobolev inequalities and generation of chaos
Matthew Rosenzweig, Sylvia Serfaty

TL;DR
This paper links the generation of chaos in particle systems to modulated logarithmic Sobolev inequalities, providing a new approach to prove exponential convergence to equilibrium in mean-field dynamics.
Contribution
It introduces the concept of modulated logarithmic Sobolev inequalities and demonstrates their role in establishing chaos generation for particle systems.
Findings
Uniform modulated logarithmic Sobolev inequalities can be established in one-dimensional cases.
The method applies to systems with singular interactions like Riesz kernels.
Generation of chaos can be proved via these inequalities under certain conditions.
Abstract
We consider mean-field limits for overdamped Langevin dynamics of particles with possibly singular interactions. It has been shown that a modulated free energy method can be used to prove the mean-field convergence or propagation of chaos for a certain class of interactions, including Riesz kernels. We show here that generation of chaos, i.e. exponential-in-time convergence to a tensorized (or iid) state starting from a nontensorized one, can be deduced from the modulated free energy method provided a uniform-in- "modulated logarithmic Sobolev inequality" holds. Proving such an inequality is a question of independent interest, which is generally difficult. As an illustration, we show that uniform modulated logarithmic Sobolev inequalities can be proven for a class of situations in one dimension.
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.
Taxonomy
TopicsMarkov Chains and Monte Carlo Methods · Quantum chaos and dynamical systems · Protein Structure and Dynamics
