Distribution (differential geometry)

About: Distribution (differential geometry) is a research topic. Over the lifetime, 911 publications have been published within this topic receiving 10149 citations.

Integrals of equivariant forms over non-compact symplectic manifolds

Abstract: This article is a result of the AIM workshop on Moment Maps and Surjectivity in Various Geometries (August 9 - 13, 2004) organized by T.Holm, E.Lerman and S.Tolman. At that workshop I was introduced to the work of T.Hausel and N.Proudfoot on hyperkahler quotients [HP]. One interesting feature of their article is that they consider integrals of equivariant forms over non-compact symplectic manifolds which do not converge, so they formally {\em define} these integrals as sums over the zeroes of vector fields, as in the Berline-Vergne localization formula. In this article we introduce a geometric-analytic regularization technique which makes such integrals converge and utilizes the symplectic structure of the manifold. We also prove that the Berline-Vergne localization formula holds for these integrals as well. The key step here is to redefine the collection of integrals \int_M alpha(X), X \in g, as a distribution on the Lie algebra g. We expect our regularization technique to generalize to non-compact group actions extending the results of [L1,L2].

Rectangular Flows for Manifold Learning

Abstract: Normalizing flows are invertible neural networks with tractable change-of-volume terms, which allows optimization of their parameters to be efficiently performed via maximum likelihood. However, data of interest is typically assumed to live in some (often unknown) low-dimensional manifold embedded in high-dimensional ambient space. The result is a modelling mismatch since -- by construction -- the invertibility requirement implies high-dimensional support of the learned distribution. Injective flows, mapping from low- to high-dimensional space, aim to fix this discrepancy by learning distributions on manifolds, but the resulting volume-change term becomes more challenging to evaluate. Current approaches either avoid computing this term entirely using various heuristics, or assume the manifold is known beforehand and therefore are not widely applicable. Instead, we propose two methods to tractably calculate the gradient of this term with respect to the parameters of the model, relying on careful use of automatic differentiation and techniques from numerical linear algebra. Both approaches perform end-to-end nonlinear manifold learning and density estimation for data projected onto this manifold. We study the trade-offs between our proposed methods, empirically verify that we outperform approaches ignoring the volume-change term by more accurately learning manifolds and the corresponding distributions on them, and show promising results on out-of-distribution detection.

On Pseudo-Einstein Real Hypersurfaces

Abstract: Let $M$ be a real hypersurface of a complex space form $M^n(c)$, $c eq0$, $n\geq 3$. We show that the Ricci tensor $S$ of $M$ satisfies $S(X,Y)=ag(X,Y)$ for any vector fields $X$ and $Y$ on the holomorphic distribution, $a$ being a constant, if and only if $M$ is a pseudo-Einstein real hypersurface.

A tentative theory of large distance physics

Abstract: A theoretical mechanism is devised to determine the large distance physics of spacetime. It is a two dimensional nonlinear model, the lambda model, set to govern the string worldsurface to remedy the failure of string theory. The lambda model is formulated to cancel the infrared divergent effects of handles at short distance on the worldsurface. The target manifold is the manifold of background spacetimes. The coupling strength is the spacetime coupling constant. The lambda model operates at 2d distance $\Lambda^{-1}$, very much shorter than the 2d distance $\mu^{-1}$ where the worldsurface is seen. A large characteristic spacetime distance $L$ is given by $L^2=\ln(\Lambda/\mu)$. Spacetime fields of wave number up to 1/L are the local coordinates for the manifold of spacetimes. The distribution of fluctuations at 2d distances shorter than $\Lambda^{-1}$ gives the {\it a priori} measure on the target manifold, the manifold of spacetimes. If this measure concentrates at a macroscopic spacetime, then, nearby, it is a measure on the spacetime fields. The lambda model thereby constructs a spacetime quantum field theory, cutoff at ultraviolet distance $L$, describing physics at distances larger than $L$. The lambda model also constructs an effective string theory with infrared cutoff $L$, describing physics at distances smaller than $L$. The lambda model evolves outward from zero 2d distance, $\Lambda^{-1} = 0$, building spacetime physics starting from $L=\infty$ and proceeding downward in $L$. $L$ can be taken smaller than any distance practical for experiments, so the lambda model, if right, gives all actually observable physics. The harmonic surfaces in the manifold of spacetimes are expected to have novel nonperturbative effects at large distances.

Contact CR-submanifolds of an indefinite Lorentzian para-Sasakian manifold

Abstract: In this paper we prove some properties of the indefinite Lorentzian para-Sasakian manifolds. Section 1 is introductory. In Sec- tion 2 we define D-totally geodesic and D ? -totally geodesic contact CR- submanifolds of an indefinite Lorentzian para-Sasakian manifold and de- duce some results concerning such a manifold. In Section 3 we state and prove some results on mixed totally geodesic contact CR-submanifolds of an indefinite Lorentzian para-Sasakian manifold. Finally, in Section 4 we obtain a result on the anti-invariant distribution of totally umbilic contact CR-submanifolds of an indefinite Lorentzian para-Sasakian man- ifold.