Propagation of chaos, Wasserstein gradient flows and toric Kahler-Einstein metrics
Robert J. Berman, Magnus Onnheim

TL;DR
This paper develops a probabilistic framework linking particle systems, Wasserstein gradient flows, and Kahler-Einstein metrics, revealing new evolution equations and a real analog of the Yau-Tian-Donaldson conjecture in complex geometry.
Contribution
It introduces a novel large N-limit evolution equation derived from stochastic particle systems, connecting Kahler geometry with Wasserstein gradient flows and extending the Yau-Tian-Donaldson conjecture.
Findings
Derivation of a drift-diffusion PDE coupled with Monge-Ampere operator
Identification of the PDE as a Wasserstein gradient flow of the K-energy functional
Establishment of a probabilistic analog of the Yau-Tian-Donaldson conjecture
Abstract
Motivated by a probabilistic approach to Kahler-Einstein metrics we consider a general non-equilibrium statistical mechanics model in Euclidean space consisting of the stochastic gradient flow of a given (possibly singular) quasi-convex N-particle interaction energy. We show that a deterministic "macroscopic" evolution equation emerges in the large N-limit of many particles. This is a strengthening of previous results which required a uniform two-sided bound on the Hessian of the interaction energy. The proof uses the theory of weak gradient flows on the Wasserstein space. Applied to the setting of permanental point processes at "negative temperature" the corresponding limiting evolution equation yields a drift-diffusion equation, coupled to the Monge-Ampere operator, whose static solutions correspond to toric Kahler-Einstein metrics. This drift-diffusion equation is the gradient flow…
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.
