Source author record

Mayukh Mukherjee

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

math.AP math.DG math.FA Machine Learning math-ph math.CV math.DS math.GN math.MG math.MP math.PR math.SP math.ST Statistics Theory

Catalog footprint

Research overview

7works

14topics

4close collaborators

Actions

Is this your work?

Open graph Browse papers

preprint2022arXiv

Heat profile, level sets and hot spots of Laplace eigenfunctions

We use probabilistic tools based on Brownian motion and Feynman-Kac formulae to investigate the heat profile for the ground state Dirichlet and second Neumann eigenfunctions. Among other topics, we comment on supremum norm bounds for ground state Dirichlet eigenfunctions and look at the corresponding Neumann problem, namely the comparison of maximum temperatures on the interior and the boundary, the latter being partially motivated by the hot spots problem. We also investigate the proximity/distance of level sets of ground state Dirichlet eigenfunctions, some with analogous statements for Neumann eigenfunctions. Domains with bottlenecks make occasional appearances as an illuminating example as well as testing ground for our theory.

preprint2021arXiv

Heating up decision boundaries: isocapacitory saturation, adversarial scenarios and generalization bounds

In the present work we study classifiers' decision boundaries via Brownian motion processes in ambient data space and associated probabilistic techniques. Intuitively, our ideas correspond to placing a heat source at the decision boundary and observing how effectively the sample points warm up. We are largely motivated by the search for a soft measure that sheds further light on the decision boundary's geometry. En route, we bridge aspects of potential theory and geometric analysis (Mazya, 2011, Grigoryan-Saloff-Coste, 2002) with active fields of ML research such as adversarial examples and generalization bounds. First, we focus on the geometric behavior of decision boundaries in the light of adversarial attack/defense mechanisms. Experimentally, we observe a certain capacitory trend over different adversarial defense strategies: decision boundaries locally become flatter as measured by isoperimetric inequalities (Ford et al, 2019); however, our more sensitive heat-diffusion metrics extend this analysis and further reveal that some non-trivial geometry invisible to plain distance-based methods is still preserved. Intuitively, we provide evidence that the decision boundaries nevertheless retain many persistent "wiggly and fuzzy" regions on a finer scale. Second, we show how Brownian hitting probabilities translate to soft generalization bounds which are in turn connected to compression and noise stability (Arora et al, 2018), and these bounds are significantly stronger if the decision boundary has controlled geometric features.

preprint2020arXiv

Polyhedral billiards, eigenfunction concentration and almost periodic control

We study dynamical properties of the billiard flow on convex polyhedra away from a neighbourhood of the non-smooth part of the boundary, called ``pockets''. We prove there are only finitely many immersed periodic tubes missing the pockets and moreover establish a new quantitative estimate for the lengths of such tubes. This extends well-known results in dimension $2$ . We then apply these dynamical results to prove a quantitative Laplace eigenfunction mass concentration near the pockets of convex polyhedral billiards. As a technical tool for proving our concentration results on irrational polyhedra, we establish a control-theoretic estimate on a product space with an almost-periodic boundary condition. This extends previously known control estimates for periodic boundary conditions, and seems to be of independent interest.

preprint2016arXiv

Discrete $z$ -filters and rings of analytic functions

Consider rings of single variable real analytic or complex entire functions, denoted by $\mathbb{K}\langle z\rangle$ . We study "discrete $z$ -filters" on $\mathbb{K}$ and their connections with the space of maximal ideals of $\mathbb{K}\langle z\rangle$ , which we characterize as a compact $T_1$ space $θ\mathbb{K}$ of discrete $z$ -ultrafilters on $\mathbb{K}$ . We show that $θ\mathbb{K}$ is a bijective continuous image of $β\mathbb{K} \setminus Q(\mathbb{K})$ , where $Q(\mathbb{K})$ is the set of far points of $β\mathbb{K}$ . $θ\mathbb{K}$ turns out to be the Wallman compactification of the canonically embedded image of $\mathbb{K}$ inside $θ\mathbb{K}$ . Using our characterization of $θ\mathbb{K}$ , we derive a Gelfand-Kolmogorov characterization of maximal ideals of $\mathbb{K}\langle z\rangle$ and show that the Krull dimension of $\mathbb{K}\langle z\rangle$ is at least $c$ . We also establish the existence of a chain of prime $z$ -filters on $\mathbb{K}$ consisting of at least $2^c$ many elements.

preprint2016arXiv

Variation of Laplace spectra of compact "nearly" hyperbolic surfaces

We use the time real analyticity of Ricci flow proved by Kotschwar to extend a result in ~\cite{B}, namely, we prove that the Laplace spectra of negatively curved compact surfaces having same genus $γ\geq 2$ , same area and same curvature bounds vary in a "controlled way", of which we give a quantitative estimate (Theorem 1.1 below). We also observe how said real analyticity can lead to unexpected conclusions about spectral properties of generic metrics on a compact surface of genus $γ\geq 2$ (Proposition 1.5 below).

preprint2015arXiv

Extremal values of the (fractional) Weinstein functional on the hyperbolic space

We make a study of Weinstein functionals, first defined in ~\cite{W}, on the hyperbolic space $\mathbb{H}^n$ . We are primarily interested in the existence of Weinstein functional maximisers, or, in other words, existence of extremal functions for the best constant of the Gagliardo-Nirenberg inequality. The main result is that the maximum value of the Weinstein functional on $\mathbb{H}^n$ is the same as that on $\mathbb{R}^n$ and the related fact that the maximum value of the Weinstein functional is not attained on $\mathbb{H}^n$ , when maximisation is done in the Sobolev space $H^1(\mathbb{H}^n)$ . This proves a conjecture made in ~\cite{CMMT} and also answers questions raised in several other papers (see, for example, ~\cite{B}). We also prove that a corresponding version of the conjecture will hold for the Weinstein functional with the fractional Laplacian as well.

preprint2015arXiv

Nonlinear travelling waves on non-Euclidean spaces

We study travelling wave solutions, that is, solutions of the form $v(t, x) = e^{iλt}u(g(t)x)$ , to nonlinear Schrödinger and Klein-Gordon equations on Riemannian manifolds, both compact and non-compact ones, with emphasis on the NLKG. Here $g(t)$ represents a one-parameter family of isometries generated by a Killing field $X$ and a case of particular interest is when $X$ has length $\leq 1$ , which leads in certain settings to hypoelliptic operators with loss of at least one derivative. In the compact case, we establish existence of travelling wave solutions via "energy" minimization methods and prove that at least compact isotropic manifolds have \emph{genuinely} travelling waves. We establish certain sharp regularity estimates on low dimensional spheres that improve results in ~\cite{T1} and carry out the subelliptic analysis for NLKG on spheres of higher dimensions utilizing their homogeneous coset space properties. Such subelliptic phenomenon have no parallel in the setting of flat spaces. We will also study related phenomenon on complete noncompact manifolds with certain symmetry assumptions using concentration-compactness type arguments.

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Mayukh Mukherjee

Research overview

Is this your work?

7 publications

Heat profile, level sets and hot spots of Laplace eigenfunctions

Heating up decision boundaries: isocapacitory saturation, adversarial scenarios and generalization bounds

Polyhedral billiards, eigenfunction concentration and almost periodic control

Discrete $z$ -filters and rings of analytic functions

Variation of Laplace spectra of compact "nearly" hyperbolic surfaces

Extremal values of the (fractional) Weinstein functional on the hyperbolic space

Nonlinear travelling waves on non-Euclidean spaces

See the researcher in context

Building this map preview

Mayukh Mukherjee

Research overview

Is this your work?

7 publications

Heat profile, level sets and hot spots of Laplace eigenfunctions

Heating up decision boundaries: isocapacitory saturation, adversarial scenarios and generalization bounds

Polyhedral billiards, eigenfunction concentration and almost periodic control

Discrete zzz-filters and rings of analytic functions

Variation of Laplace spectra of compact "nearly" hyperbolic surfaces

Extremal values of the (fractional) Weinstein functional on the hyperbolic space

Nonlinear travelling waves on non-Euclidean spaces

See the researcher in context

Building this map preview

Discrete $z$ -filters and rings of analytic functions