Showing 1–50 of 65 results for author: Bernstein, J

Old Optimizer, New Norm: An Anthology

Authors: Jeremy Bernstein, Laker Newhouse

Abstract: Deep learning optimizers are often motivated through a mix of convex and approximate second-order theory. We select three such methods -- Adam, Shampoo and Prodigy -- and argue that each method can instead be understood as a squarely first-order method without convexity assumptions. In fact, after switching off exponential moving averages, each method is equivalent to steepest descent under a part… ▽ More Deep learning optimizers are often motivated through a mix of convex and approximate second-order theory. We select three such methods -- Adam, Shampoo and Prodigy -- and argue that each method can instead be understood as a squarely first-order method without convexity assumptions. In fact, after switching off exponential moving averages, each method is equivalent to steepest descent under a particular norm. By generalizing this observation, we chart a new design space for training algorithms. Different operator norms should be assigned to different tensors based on the role that the tensor plays within the network. For example, while linear and embedding layers may have the same weight space of $\mathbb{R}^{m\times n}$, these layers play different roles and should be assigned different norms. We hope that this idea of carefully metrizing the neural architecture might lead to more stable, scalable and indeed faster training. △ Less

Submitted 6 December, 2024; v1 submitted 30 September, 2024; originally announced September 2024.

Li-Yau Conformal Volume and Colding-Minicozzi Entropy of Self-Shrinkers

Authors: Jacob Bernstein

Abstract: We show the (normalized) Li-Yau conformal volume of a self-shrinker of mean curvature flow in Euclidean space bounds its Colding-Minicozzi entropy from below. This bound is independent of codimension and sharp on planes. As an application we verify a conjecture of Colding-Minicozzi about the entropy of closed self-shrinkers of arbitrary codimension for self-shrinkers that are topologically two-dim… ▽ More We show the (normalized) Li-Yau conformal volume of a self-shrinker of mean curvature flow in Euclidean space bounds its Colding-Minicozzi entropy from below. This bound is independent of codimension and sharp on planes. As an application we verify a conjecture of Colding-Minicozzi about the entropy of closed self-shrinkers of arbitrary codimension for self-shrinkers that are topologically two-dimensional real projective planes. As part of the proof we introduce two auxiliary functionals which we call stable conformal volume and virtual entropy which should be of independent interest. △ Less

Submitted 2 July, 2024; originally announced July 2024.

Comments: 28 pages

MSC Class: 53E10; 53C18; 53C42

Existence of monotone Morse flow lines of the expander functional

Authors: Jacob Bernstein, Letian Chen, Lu Wang

Abstract: Given a smooth asymptotically conical self-expander that is strictly unstable we construct a (singular) Morse flow line of the expander functional that connects it to a stable self-expander. This flow is monotone in a suitable sense and has small singular set. Given a smooth asymptotically conical self-expander that is strictly unstable we construct a (singular) Morse flow line of the expander functional that connects it to a stable self-expander. This flow is monotone in a suitable sense and has small singular set. △ Less

Submitted 12 April, 2024; originally announced April 2024.

Comments: 46 pages

MSC Class: 53E10; 49Q20

Lower Bounds on Density for Topologically Nontrivial Minimal Cones up to Dimension Six

Authors: Jacob Bernstein, Lu Wang

Abstract: We prove lower bounds on the density of regular minimal cones of dimension less than seven provided the complements of the cones are topologically nontrivial. We prove lower bounds on the density of regular minimal cones of dimension less than seven provided the complements of the cones are topologically nontrivial. △ Less

Submitted 5 April, 2024; originally announced April 2024.

Comments: 19 pages; 1 figure

MSC Class: 53A10; 53E10

Hypergeometric orthogonal polynomials of Jacobi type

Authors: Joseph Bernstein, Dmitry Gourevitch, Siddhartha Sahi

Abstract: Motivated by the theory of hypergeometric orthogonal polynomials, we consider quasi-orthogonal polynomial families - those that are orthogonal with respect to a non-degenerate bilinear form defined by a linear functional - in which the ratio of successive coefficients is given by a rational function $f(u,s)$ which is polynomial in $u$. We call this a family of Jacobi type. Our main result is tha… ▽ More Motivated by the theory of hypergeometric orthogonal polynomials, we consider quasi-orthogonal polynomial families - those that are orthogonal with respect to a non-degenerate bilinear form defined by a linear functional - in which the ratio of successive coefficients is given by a rational function $f(u,s)$ which is polynomial in $u$. We call this a family of Jacobi type. Our main result is that, up to rescaling and renormalization, there are only five families of Jacobi type. These are the classical families of Jacobi, Laguerre and Bessel polynomials, and two more one parameter families $E_n^{(c)},F_n^{(c)}$. Each family arises as a specialization of some hypergeometric series. The last two families can also be expressed through Lommel polynomials, and they are orthogonal with respect to a positive measure on $\mathbb{R}$ for $c>0$ and $c>-1$ respectively. We also consider the more general rational families, i.e. quasi-orthogonal families in which the ratio $f(u,s)$ of successive coefficients is allowed to be rational in $u$ as well. △ Less

Submitted 4 November, 2024; v1 submitted 26 January, 2024; originally announced January 2024.

Comments: 24 p, most info in the first 5 pages. v3: minor corrections

MSC Class: 33C45; 33D45

The CR-Volume of Horizontal Submanifolds of Spheres

Authors: Jacob Bernstein, Arunima Bhattacharya

Abstract: We study an analog in CR-geometry of the conformal volume of Li-Yau. In particular, to submanifolds of odd-dimensional spheres that are Legendrian or, more generally, horizontal with respect to the sphere's standard CR-structure we associate a quantity that is invariant under the CR-automorphisms of the sphere. We apply this concept to a corresponding notion of Willmore energy. We study an analog in CR-geometry of the conformal volume of Li-Yau. In particular, to submanifolds of odd-dimensional spheres that are Legendrian or, more generally, horizontal with respect to the sphere's standard CR-structure we associate a quantity that is invariant under the CR-automorphisms of the sphere. We apply this concept to a corresponding notion of Willmore energy. △ Less

Submitted 20 January, 2024; originally announced January 2024.

Comments: 38 pages

Minimal surfaces and Colding-Minicozzi entropy in complex hyperbolic space

Authors: Jacob Bernstein, Arunima Bhattacharya

Abstract: We study notions of asymptotic regularity for a class of minimal submanifolds of complex hyperbolic space that includes minimal Lagrangian submanifolds. As an application, we show a relationship between an appropriate formulation of Colding-Minicozzi entropy and a quantity we call the CR-volume that is computed from the asymptotic geometry of such submanifolds. We study notions of asymptotic regularity for a class of minimal submanifolds of complex hyperbolic space that includes minimal Lagrangian submanifolds. As an application, we show a relationship between an appropriate formulation of Colding-Minicozzi entropy and a quantity we call the CR-volume that is computed from the asymptotic geometry of such submanifolds. △ Less

Submitted 16 July, 2023; originally announced July 2023.

Comments: 36 pages

MSC Class: 53A10; 32Q05

Rigidity Properties of Colding-Minicozzi Entropies

Authors: Jacob Bernstein

Abstract: We show certain rigidity for minimizers of generalized Colding-Minicozzi entropies. The proofs are elementary and work even in situations where the generalized entropies are not monotone along mean curvature flow. We show certain rigidity for minimizers of generalized Colding-Minicozzi entropies. The proofs are elementary and work even in situations where the generalized entropies are not monotone along mean curvature flow. △ Less

Submitted 20 June, 2023; originally announced June 2023.

Comments: 14 pages

MSC Class: 53A07; 53A10

An Analysis of the Johnson-Lindenstrauss Lemma with the Bivariate Gamma Distribution

Authors: Jason Bernstein, Alec M. Dunton, Benjamin W. Priest

Abstract: Probabilistic proofs of the Johnson-Lindenstrauss lemma imply that random projection can reduce the dimension of a data set and approximately preserve pairwise distances. If a distance being approximately preserved is called a success, and the complement of this event is called a failure, then such a random projection likely results in no failures. Assuming a Gaussian random projection, the lemma… ▽ More Probabilistic proofs of the Johnson-Lindenstrauss lemma imply that random projection can reduce the dimension of a data set and approximately preserve pairwise distances. If a distance being approximately preserved is called a success, and the complement of this event is called a failure, then such a random projection likely results in no failures. Assuming a Gaussian random projection, the lemma is proved by showing that the no-failure probability is positive using a combination of Bonferroni's inequality and Markov's inequality. This paper modifies this proof in two ways to obtain a greater lower bound on the no-failure probability. First, Bonferroni's inequality is applied to pairs of failures instead of individual failures. Second, since a pair of projection errors has a bivariate gamma distribution, the probability of a pair of successes is bounded using an inequality from Jensen (1969). If $n$ is the number of points to be embedded and $μ$ is the probability of a success, then this leads to an increase in the lower bound on the no-failure probability of $\frac{1}{2}\binom{n}{2}(1-μ)^2$ if $\binom{n}{2}$ is even and $\frac{1}{2}\left(\binom{n}{2}-1\right)(1-μ)^2$ if $\binom{n}{2}$ is odd. For example, if $n=10^5$ points are to be embedded in $k=10^4$ dimensions with a tolerance of $ε=0.1$, then the improvement in the lower bound is on the order of $10^{-14}$. We also show that further improvement is possible if the inequality in Jensen (1969) extends to three successes, though we do not have a proof of this result. △ Less

Submitted 12 July, 2024; v1 submitted 26 May, 2023; originally announced May 2023.

Comments: 20 pages, 5 figures. This revision improves figure 2 and format of citations

Report number: LLNL-TR-844277

Automatic Gradient Descent: Deep Learning without Hyperparameters

Authors: Jeremy Bernstein, Chris Mingard, Kevin Huang, Navid Azizan, Yisong Yue

Abstract: The architecture of a deep neural network is defined explicitly in terms of the number of layers, the width of each layer and the general network topology. Existing optimisation frameworks neglect this information in favour of implicit architectural information (e.g. second-order methods) or architecture-agnostic distance functions (e.g. mirror descent). Meanwhile, the most popular optimiser in pr… ▽ More The architecture of a deep neural network is defined explicitly in terms of the number of layers, the width of each layer and the general network topology. Existing optimisation frameworks neglect this information in favour of implicit architectural information (e.g. second-order methods) or architecture-agnostic distance functions (e.g. mirror descent). Meanwhile, the most popular optimiser in practice, Adam, is based on heuristics. This paper builds a new framework for deriving optimisation algorithms that explicitly leverage neural architecture. The theory extends mirror descent to non-convex composite objective functions: the idea is to transform a Bregman divergence to account for the non-linear structure of neural architecture. Working through the details for deep fully-connected networks yields automatic gradient descent: a first-order optimiser without any hyperparameters. Automatic gradient descent trains both fully-connected and convolutional networks out-of-the-box and at ImageNet scale. A PyTorch implementation is available at https://github.com/jxbz/agd and also in Appendix B. Overall, the paper supplies a rigorous theoretical foundation for a next-generation of architecture-dependent optimisers that work automatically and without hyperparameters. △ Less

Submitted 11 April, 2023; originally announced April 2023.

Colding-Minicozzi Entropies in Cartan-Hadamard Manifolds

Authors: Jacob Bernstein, Arunima Bhattacharya

Abstract: We introduce a family of functionals on submanifolds of Cartan-Hadamard manifolds that generalize the Colding-Minicozzi entropy of submanifolds of Euclidean space. We show that these functionals are monotone under mean curvature flow under natural conditions. As a consequence, we obtain sharp lower bounds on these entropies for certain closed hypersurfaces and observe a novel rigidity phenomenon. We introduce a family of functionals on submanifolds of Cartan-Hadamard manifolds that generalize the Colding-Minicozzi entropy of submanifolds of Euclidean space. We show that these functionals are monotone under mean curvature flow under natural conditions. As a consequence, we obtain sharp lower bounds on these entropies for certain closed hypersurfaces and observe a novel rigidity phenomenon. △ Less

Submitted 25 November, 2022; originally announced November 2022.

Comments: 34 pages

MSC Class: 53C44

Optimisation & Generalisation in Networks of Neurons

Authors: Jeremy Bernstein

Abstract: The goal of this thesis is to develop the optimisation and generalisation theoretic foundations of learning in artificial neural networks. On optimisation, a new theoretical framework is proposed for deriving architecture-dependent first-order optimisation algorithms. The approach works by combining a "functional majorisation" of the loss function with "architectural perturbation bounds" that enco… ▽ More The goal of this thesis is to develop the optimisation and generalisation theoretic foundations of learning in artificial neural networks. On optimisation, a new theoretical framework is proposed for deriving architecture-dependent first-order optimisation algorithms. The approach works by combining a "functional majorisation" of the loss function with "architectural perturbation bounds" that encode an explicit dependence on neural architecture. The framework yields optimisation methods that transfer hyperparameters across learning problems. On generalisation, a new correspondence is proposed between ensembles of networks and individual networks. It is argued that, as network width and normalised margin are taken large, the space of networks that interpolate a particular training set concentrates on an aggregated Bayesian method known as a "Bayes point machine". This correspondence provides a route for transferring PAC-Bayesian generalisation theorems over to individual networks. More broadly, the correspondence presents a fresh perspective on the role of regularisation in networks with vastly more parameters than data. △ Less

Submitted 18 October, 2022; originally announced October 2022.

Comments: PhD thesis

$P$-strict promotion and $Q$-partition rowmotion: the graded case

Authors: Joseph Bernstein, Jessica Striker, Corey Vorland

Abstract: Promotion and rowmotion are intriguing actions in dynamical algebraic combinatorics which have inspired much work in recent years. In this paper, we study $P$-strict labelings of a finite, graded poset $P$ of rank $n$ and labels at most $q$, which generalize semistandard Young tableaux with $n$ rows and entries at most $q$, under promotion. These $P$-strict labelings are in equivariant bijection w… ▽ More Promotion and rowmotion are intriguing actions in dynamical algebraic combinatorics which have inspired much work in recent years. In this paper, we study $P$-strict labelings of a finite, graded poset $P$ of rank $n$ and labels at most $q$, which generalize semistandard Young tableaux with $n$ rows and entries at most $q$, under promotion. These $P$-strict labelings are in equivariant bijection with $Q$-partitions under rowmotion, where $Q$ equals the product of $P$ and a chain of $q-n-1$ elements. We study the case where $P$ equals the product of chains in detail, yielding new homomesy and order results in the realm of tableaux and beyond. Furthermore, we apply the bijection to the cases in which $P$ is a minuscule poset and when $P$ is the three element $V$ poset. Finally, we give resonance results for promotion on $P$-strict labelings and rowmotion on $Q$-partitions. △ Less

Submitted 11 July, 2023; v1 submitted 10 May, 2022; originally announced May 2022.

Comments: 24 pages, 7 figures

MSC Class: 05E18

Journal ref: European Journal of Combinatorics 115 (2024), no. 103776, 24 pp

Strong density of spherical characters attached to unipotent subgroups

Authors: Avraham Aizenbud, Joseph Bernstein, Eitan Sayag

Abstract: We prove the following result in relative representation theory of a reductive p-adic group $G$: Let $U$ be the unipotent radical of a minimal parabolic subgroup of $G$, and let $ψ$ be an arbitrary smooth character of $U$. Let $S \subset Irr(G)$ be a Zariski dense collection of irreducible representations of $G$. Then the span of the Bessel distributions $B_π$ attached to representations $π$ fro… ▽ More We prove the following result in relative representation theory of a reductive p-adic group $G$: Let $U$ be the unipotent radical of a minimal parabolic subgroup of $G$, and let $ψ$ be an arbitrary smooth character of $U$. Let $S \subset Irr(G)$ be a Zariski dense collection of irreducible representations of $G$. Then the span of the Bessel distributions $B_π$ attached to representations $π$ from $S$ is dense in the space $\mathcal S^*(G)^{U\times U,ψ\times ψ}$ of all $(U\times U,ψ\times ψ)$-equivariant distributions on $G.$ We base our proof on the following results: 1. The category of smooth representations $\mathcal M(G)$ is Cohen-Macaulay. 2. The module $ind_U^G(ψ)$ is a projective module. △ Less

Submitted 10 February, 2022; originally announced February 2022.

Comments: 17 pages

MSC Class: 20G05; 20G25; 46F99; 43A85

A Sharp Isoperimetric Property of the Renormalized Area of a Minimal Surface in Hyperbolic Space

Authors: Jacob Bernstein

Abstract: We prove an inequality bounding the renormalized area of a complete minimal surface in hyperbolic space in terms of the conformal length of its ideal boundary. We prove an inequality bounding the renormalized area of a complete minimal surface in hyperbolic space in terms of the conformal length of its ideal boundary. △ Less

Submitted 27 April, 2021; originally announced April 2021.

Comments: 16 pages

MSC Class: 53A10

$P$-strict promotion and $B$-bounded rowmotion, with applications to tableaux of many flavors

Authors: Joseph Bernstein, Jessica Striker, Corey Vorland

Abstract: We define P-strict labelings for a finite poset P as a generalization of semistandard Young tableaux and show that promotion on these objects is in equivariant bijection with a toggle action on B-bounded Q-partitions of an associated poset Q. In many nice cases, this toggle action is conjugate to rowmotion. We apply this result to flagged tableaux, Gelfand-Tsetlin patterns, and symplectic tableaux… ▽ More We define P-strict labelings for a finite poset P as a generalization of semistandard Young tableaux and show that promotion on these objects is in equivariant bijection with a toggle action on B-bounded Q-partitions of an associated poset Q. In many nice cases, this toggle action is conjugate to rowmotion. We apply this result to flagged tableaux, Gelfand-Tsetlin patterns, and symplectic tableaux, obtaining new cyclic sieving and homomesy conjectures. We also show P-strict promotion can be equivalently defined using Bender-Knuth and jeu de taquin perspectives. △ Less

Submitted 4 August, 2021; v1 submitted 22 December, 2020; originally announced December 2020.

Comments: 39 pages, 14 figures

MSC Class: 05A19; 05E18

Journal ref: Combinatorial Theory, Volume 1, 2021

Colding Minicozzi Entropy in Hyperbolic Space

Authors: Jacob Bernstein

Abstract: This note introduces a notion of entropy for submanifolds of hyperbolic space analogous to the one introduced by Colding and Minicozzi for submanifolds of Euclidean space. Several properties are proved for this quantity including monotonicity along mean curvature flow in low dimensions and a connection with the conformal volume. This note introduces a notion of entropy for submanifolds of hyperbolic space analogous to the one introduced by Colding and Minicozzi for submanifolds of Euclidean space. Several properties are proved for this quantity including monotonicity along mean curvature flow in low dimensions and a connection with the conformal volume. △ Less

Submitted 20 July, 2020; originally announced July 2020.

Comments: 16 pages

MSC Class: 53C44

Learning compositional functions via multiplicative weight updates

Authors: Jeremy Bernstein, Jiawei Zhao, Markus Meister, Ming-Yu Liu, Anima Anandkumar, Yisong Yue

Abstract: Compositionality is a basic structural feature of both biological and artificial neural networks. Learning compositional functions via gradient descent incurs well known problems like vanishing and exploding gradients, making careful learning rate tuning essential for real-world applications. This paper proves that multiplicative weight updates satisfy a descent lemma tailored to compositional fun… ▽ More Compositionality is a basic structural feature of both biological and artificial neural networks. Learning compositional functions via gradient descent incurs well known problems like vanishing and exploding gradients, making careful learning rate tuning essential for real-world applications. This paper proves that multiplicative weight updates satisfy a descent lemma tailored to compositional functions. Based on this lemma, we derive Madam -- a multiplicative version of the Adam optimiser -- and show that it can train state of the art neural network architectures without learning rate tuning. We further show that Madam is easily adapted to train natively compressed neural networks by representing their weights in a logarithmic number system. We conclude by drawing connections between multiplicative weight updates and recent findings about synapses in biology. △ Less

Submitted 8 January, 2021; v1 submitted 25 June, 2020; originally announced June 2020.

Hidden sign in Langlands' correspondence

Authors: Joseph Bernstein

Abstract: In this note I describe some modification of the Langlands correspondence and explain why it is more natural. I will also discuss its significance to the theory of automorphic L-functions. In this note I describe some modification of the Langlands correspondence and explain why it is more natural. I will also discuss its significance to the theory of automorphic L-functions. △ Less

Submitted 19 May, 2020; v1 submitted 22 April, 2020; originally announced April 2020.

Comments: Version 2 corrects some errors and adds some bibliography

MSC Class: 20G05

Closed hypersurfaces of low entropy in $\mathbb{R}^4$ are isotopically trivial

Authors: Jacob Bernstein, Lu Wang

Abstract: We show that any closed connected hypersurface in $\mathbb{R}^4$ with entropy less than or equal to that of the round cylinder is smoothly isotopic to the standard three-sphere. We show that any closed connected hypersurface in $\mathbb{R}^4$ with entropy less than or equal to that of the round cylinder is smoothly isotopic to the standard three-sphere. △ Less

Submitted 30 March, 2020; originally announced March 2020.

Comments: 18 pages, 0 figure

MSC Class: 53C44

A mountain-pass theorem for asymptotically conical self-expanders

Authors: Jacob Bernstein, Lu Wang

Abstract: We develop a min-max theory for asymptotically conical self-expanders of mean curvature flow. In particular, we show that given two distinct strictly stable self-expanders that are asymptotic to the same cone and bound a domain, there exists a new asymptotically conical self-expander trapped between the two. We develop a min-max theory for asymptotically conical self-expanders of mean curvature flow. In particular, we show that given two distinct strictly stable self-expanders that are asymptotic to the same cone and bound a domain, there exists a new asymptotically conical self-expander trapped between the two. △ Less

Submitted 30 March, 2020; originally announced March 2020.

Comments: 49 pages, 0 figure

MSC Class: 53C44

Symmetry and Rigidity of Minimal Surfaces with Plateau-like Singularities

Authors: Jacob Bernstein, Francesco Maggi

Abstract: By employing the method of moving planes in a novel way we extend some classical symmetry and rigidity results for smooth minimal surfaces to surfaces that have singularities of the sort typically observed in soap films. By employing the method of moving planes in a novel way we extend some classical symmetry and rigidity results for smooth minimal surfaces to surfaces that have singularities of the sort typically observed in soap films. △ Less

Submitted 5 October, 2020; v1 submitted 3 March, 2020; originally announced March 2020.

Comments: 30 pages; an appendix about a maximum principle for geodesic nets has been added

On the distance between two neural networks and the stability of learning

Authors: Jeremy Bernstein, Arash Vahdat, Yisong Yue, Ming-Yu Liu

Abstract: This paper relates parameter distance to gradient breakdown for a broad class of nonlinear compositional functions. The analysis leads to a new distance function called deep relative trust and a descent lemma for neural networks. Since the resulting learning rule seems to require little to no learning rate tuning, it may unlock a simpler workflow for training deeper and more complex neural network… ▽ More This paper relates parameter distance to gradient breakdown for a broad class of nonlinear compositional functions. The analysis leads to a new distance function called deep relative trust and a descent lemma for neural networks. Since the resulting learning rule seems to require little to no learning rate tuning, it may unlock a simpler workflow for training deeper and more complex neural networks. The Python code used in this paper is here: https://github.com/jxbz/fromage. △ Less

Submitted 8 January, 2021; v1 submitted 9 February, 2020; originally announced February 2020.

On the meromorphic continuation of Eisenstein series

Authors: Joseph Bernstein, Erez Lapid

Abstract: Eisenstein series are ubiquitous in the theory of automorphic forms. The traditional proofs of the meromorphic continuation of Eisenstein series, due to Selberg and Langlands, start with cuspidal Eisenstein series as a special case, and deduce the general case from spectral theory. We present a "soft" proof which relies only on rudimentary Fredholm theory (needed only in the number field case).… ▽ More Eisenstein series are ubiquitous in the theory of automorphic forms. The traditional proofs of the meromorphic continuation of Eisenstein series, due to Selberg and Langlands, start with cuspidal Eisenstein series as a special case, and deduce the general case from spectral theory. We present a "soft" proof which relies only on rudimentary Fredholm theory (needed only in the number field case). It is valid for Eisenstein series induced from an arbitrary automorphic form. The proof relies on the principle of meromorphic continuation. It is close in spirit to Selberg's later proofs. △ Less

Submitted 7 February, 2022; v1 submitted 6 November, 2019; originally announced November 2019.

Comments: Simplified proof of local finiteness

Relative expander entropy in the presence of a two-sided obstacle and applications

Authors: Jacob Bernstein, Lu Wang

Abstract: We study a notion of relative entropy motivated by self-expanders of mean curvature flow. In particular, we obtain the existence of this quantity for arbitrary hypersurfaces trapped between two disjoint self-expanders asymptotic to the same cone. This allows us to begin to develop the variational theory for the relative entropy functional for the associated obstacle problem. We also obtain a versi… ▽ More We study a notion of relative entropy motivated by self-expanders of mean curvature flow. In particular, we obtain the existence of this quantity for arbitrary hypersurfaces trapped between two disjoint self-expanders asymptotic to the same cone. This allows us to begin to develop the variational theory for the relative entropy functional for the associated obstacle problem. We also obtain a version of the forward monotonicity formula for mean curvature flow proposed by Ilmanen. △ Less

Submitted 30 March, 2020; v1 submitted 18 June, 2019; originally announced June 2019.

Comments: 34 pages, 0 figures

MSC Class: 53C44

Topological Uniqueness for Self-expanders of Small Entropy

Authors: Jacob Bernstein, Lu Wang

Abstract: For a fixed regular cone in Euclidean space with small entropy we show that all smooth self-expanding solutions of the mean curvature flow that are asymptotic to the cone are in the same isotopy class. For a fixed regular cone in Euclidean space with small entropy we show that all smooth self-expanding solutions of the mean curvature flow that are asymptotic to the cone are in the same isotopy class. △ Less

Submitted 30 March, 2020; v1 submitted 7 February, 2019; originally announced February 2019.

Comments: 33 pages, 0 fiigures

MSC Class: 53C44

An integer degree for asymptotically conical self-expanders

Authors: Jacob Bernstein, Lu Wang

Abstract: We establish the existence of an integer degree for the natural projection map from the space of parameterizations of asymptotically conical self-expanders to the space of parameterizations of the asymptotic cones when this map is proper. As an application we show that there is an open set in the space of cones in the three-dimensional Euclidean space for which each cone in the set has a strictly… ▽ More We establish the existence of an integer degree for the natural projection map from the space of parameterizations of asymptotically conical self-expanders to the space of parameterizations of the asymptotic cones when this map is proper. As an application we show that there is an open set in the space of cones in the three-dimensional Euclidean space for which each cone in the set has a strictly unstable self-expanding annuli asymptotic to it. △ Less

Submitted 17 July, 2018; originally announced July 2018.

Comments: 42 pages, 0 figures

MSC Class: 53C44

Smooth compactness for spaces of asymptotically conical self-expanders of mean curvature flow

Authors: Jacob Bernstein, Lu Wang

Abstract: We show compactness in the locally smooth topology for certain natural families of asymptotically conical self-expanding solutions of mean curvature flow. Specifically, we show such compactness for the set of all two-dimensional self-expanders of a fixed topological type and, in all dimensions, for the set of self-expanders of low entropy and for the set of mean convex self-expanders with strictly… ▽ More We show compactness in the locally smooth topology for certain natural families of asymptotically conical self-expanding solutions of mean curvature flow. Specifically, we show such compactness for the set of all two-dimensional self-expanders of a fixed topological type and, in all dimensions, for the set of self-expanders of low entropy and for the set of mean convex self-expanders with strictly mean convex asymptotic cones. From this we deduce that the natural projection map from the space of parameterizations of asymptotically conical self-expanders to the space of parameterizations of the asymptotic cones is proper for these classes. △ Less

Submitted 21 July, 2018; v1 submitted 24 April, 2018; originally announced April 2018.

Comments: 19 pages, 0 figures, fix minor typos and update references

MSC Class: 53C44

signSGD: Compressed Optimisation for Non-Convex Problems

Authors: Jeremy Bernstein, Yu-Xiang Wang, Kamyar Azizzadenesheli, Anima Anandkumar

Abstract: Training large neural networks requires distributing learning across multiple workers, where the cost of communicating gradients can be a significant bottleneck. signSGD alleviates this problem by transmitting just the sign of each minibatch stochastic gradient. We prove that it can get the best of both worlds: compressed gradients and SGD-level convergence rate. The relative $\ell_1/\ell_2$ geome… ▽ More Training large neural networks requires distributing learning across multiple workers, where the cost of communicating gradients can be a significant bottleneck. signSGD alleviates this problem by transmitting just the sign of each minibatch stochastic gradient. We prove that it can get the best of both worlds: compressed gradients and SGD-level convergence rate. The relative $\ell_1/\ell_2$ geometry of gradients, noise and curvature informs whether signSGD or SGD is theoretically better suited to a particular problem. On the practical side we find that the momentum counterpart of signSGD is able to match the accuracy and convergence speed of Adam on deep Imagenet models. We extend our theory to the distributed setting, where the parameter server uses majority vote to aggregate gradient signs from each worker enabling 1-bit compression of worker-server communication in both directions. Using a theorem by Gauss we prove that majority vote can achieve the same reduction in variance as full precision distributed SGD. Thus, there is great promise for sign-based optimisation schemes to achieve fast communication and fast convergence. Code to reproduce experiments is to be found at https://github.com/jxbz/signSGD . △ Less

Submitted 7 August, 2018; v1 submitted 12 February, 2018; originally announced February 2018.

The level set flow of a hypersurface in $\mathbb R^4$ of low entropy does not disconnect

Authors: Jacob Bernstein, Shengwen Wang

Abstract: We show that if $Σ\subset \mathbb R^4$ is a closed, connected hypersurface with entropy $λ(Σ)\leq λ(\mathbb{S}^2\times \mathbb R)$, then the level set flow of $Σ$ never disconnects. We also obtain a sharp version of the forward clearing out lemma for non-fattening flows in $\mathbb R^4$ of low entropy. We show that if $Σ\subset \mathbb R^4$ is a closed, connected hypersurface with entropy $λ(Σ)\leq λ(\mathbb{S}^2\times \mathbb R)$, then the level set flow of $Σ$ never disconnects. We also obtain a sharp version of the forward clearing out lemma for non-fattening flows in $\mathbb R^4$ of low entropy. △ Less

Submitted 29 August, 2020; v1 submitted 15 January, 2018; originally announced January 2018.

Comments: Strengthened statement of Proposition 3.4 and fixed a gap in proof of Proposition 3.5

MSC Class: 53C44

The space of asymptotically conical self-expanders of mean curvature flow

Authors: Jacob Bernstein, Lu Wang

Abstract: We show that the space of asymptotically conical self-expanders of the mean curvature flow is a smooth Banach manifold. An immediate consequence is that non-degenerate self-expanders -- that is, those self-expanders that admit no non-trivial normal Jacobi fields that fix the asymptotic cone -- are generic in a certain sense. We show that the space of asymptotically conical self-expanders of the mean curvature flow is a smooth Banach manifold. An immediate consequence is that non-degenerate self-expanders -- that is, those self-expanders that admit no non-trivial normal Jacobi fields that fix the asymptotic cone -- are generic in a certain sense. △ Less

Submitted 21 July, 2018; v1 submitted 12 December, 2017; originally announced December 2017.

Comments: 40 pages, 0 figure, fix typos and update references

MSC Class: 53C44

Asymptotic structure of almost eigenfunctions of drift Laplacians on conical ends

Authors: Jacob Bernstein

Abstract: We use a weighted variant of the frequency functions introduced by Almgren to prove sharp asymptotic estimates for almost eigenfunctions of the drift Laplacian associated to the Gaussian weight on an asymptotically conical end. As a consequence, we obtain a purely elliptic proof of a result of L. Wang on the uniqueness of self-shrinkers of the mean curvature flow asymptotic to a given cone. Anothe… ▽ More We use a weighted variant of the frequency functions introduced by Almgren to prove sharp asymptotic estimates for almost eigenfunctions of the drift Laplacian associated to the Gaussian weight on an asymptotically conical end. As a consequence, we obtain a purely elliptic proof of a result of L. Wang on the uniqueness of self-shrinkers of the mean curvature flow asymptotic to a given cone. Another consequence is a unique continuation property for self-expanders of the mean curvature flow that flow from a cone. △ Less

Submitted 12 December, 2017; v1 submitted 23 August, 2017; originally announced August 2017.

Comments: 26 pages. Fixed typos and included additional references. Weakened hypotheses of Theorem 9.1

MSC Class: 53C44; 35J15

Contractions of Representations and Algebraic Families of Harish-Chandra Modules

Authors: Joseph Bernstein, Nigel Higson, Eyal Subag

Abstract: We examine from an algebraic point of view some families of unitary group representations that arise in mathematical physics and are associated to contraction families of Lie groups. The contraction families of groups relate different real forms of a reductive group and are continuously parametrized, but the unitary representations are defined over a parameter subspace that includes both discrete… ▽ More We examine from an algebraic point of view some families of unitary group representations that arise in mathematical physics and are associated to contraction families of Lie groups. The contraction families of groups relate different real forms of a reductive group and are continuously parametrized, but the unitary representations are defined over a parameter subspace that includes both discrete and continuous parts. Both finite- and infinite-dimensional representations can occur, even within the same family. We shall study the simplest nontrivial examples, and use the concepts of algebraic families of Harish-Chandra pairs and Harish-Chandra modules, introduced in a previous paper, together with the Jantzen filtration, to construct these families of unitary representations algebraically. △ Less

Submitted 10 September, 2017; v1 submitted 11 March, 2017; originally announced March 2017.

MSC Class: 22E47; 20G; 20C; 14D; 14L

Deligne-Lusztig duality and wonderful compactification

Authors: Joseph Bernstein, Roman Bezrukavnikov, David Kazhdan

Abstract: We use geometry of the wonderful compactification to obtain a new proof of the relation between Deligne-Lusztig (or Alvis-Curtis) duality for $p$-adic groups and the homological duality. This provides a new way to introduce an involution on the set of irreducible representations of the group which has been defined by A. Zelevinsky for $G=GL(n)$ by A.-M. Aubert in general (less direct geometric app… ▽ More We use geometry of the wonderful compactification to obtain a new proof of the relation between Deligne-Lusztig (or Alvis-Curtis) duality for $p$-adic groups and the homological duality. This provides a new way to introduce an involution on the set of irreducible representations of the group which has been defined by A. Zelevinsky for $G=GL(n)$ by A.-M. Aubert in general (less direct geometric approaches to this duality have been developed earlier by Schneider-Stuhler and by the second author). As a byproduct we obtain a description of the Serre functor for representations of a p-adic group. △ Less

Submitted 10 October, 2018; v1 submitted 25 January, 2017; originally announced January 2017.

Comments: 14pp. The only change from the previous version is a modified grant acknowledgment

Algebraic Families of Harish-Chandra Pairs

Authors: Joseph Bernstein, Nigel Higson, Eyal Subag

Abstract: Mathematical physicists have studied degenerations of Lie groups and their representations, which they call contractions. In this paper we study these contractions, and also other families, within the framework of algebraic families of Harish-Chandra modules. We construct a family that incorporates both a real reductive group and its compact form, separate parts of which have been studied individu… ▽ More Mathematical physicists have studied degenerations of Lie groups and their representations, which they call contractions. In this paper we study these contractions, and also other families, within the framework of algebraic families of Harish-Chandra modules. We construct a family that incorporates both a real reductive group and its compact form, separate parts of which have been studied individually as contractions. We give a complete classification of generically irreducible families of Harish-Chandra modules in the case of the family associated to SL(2, R). △ Less

Submitted 10 September, 2017; v1 submitted 11 October, 2016; originally announced October 2016.

Comments: This is the first paper of a series of two

MSC Class: 22E47; 20G; 20C; 14D; 14L

Hausdorff Stability of the Round Two-Sphere Under Small Perturbations of the Entropy

Authors: Jacob Bernstein, Lu Wang

Abstract: We show that if a closed surface in $\mathbb{R}^3$ has entropy near to that of the unit two-sphere, then the surface is close to a round two-sphere in the Hausdorff distance. We show that if a closed surface in $\mathbb{R}^3$ has entropy near to that of the unit two-sphere, then the surface is close to a round two-sphere in the Hausdorff distance. △ Less

Submitted 8 August, 2016; originally announced August 2016.

Comments: 13 pages, 0 figure

MSC Class: 53C44

Topology of Closed Hypersurfaces of Small Entropy

Authors: Jacob Bernstein, Lu Wang

Abstract: We use a weak mean curvature flow together with a surgery procedure to show that all closed hypersurfaces in $\mathbb{R}^4$ with entropy less than or equal to that of $\mathbb{S}^2\times \mathbb{R}$, the round cylinder in $\mathbb{R}^4$, are diffeomorphic to $\mathbb{S}^3$. We use a weak mean curvature flow together with a surgery procedure to show that all closed hypersurfaces in $\mathbb{R}^4$ with entropy less than or equal to that of $\mathbb{S}^2\times \mathbb{R}$, the round cylinder in $\mathbb{R}^4$, are diffeomorphic to $\mathbb{S}^3$. △ Less

Submitted 28 June, 2016; v1 submitted 2 November, 2015; originally announced November 2015.

Comments: 24 pages. Revised version

MSC Class: 53C44

Journal ref: Geom. Topol. 22 (2018) 1109-1141

A Topological Property of Asymptotically Conical Self-Shrinkers of Small Entropy

Authors: Jacob Bernstein, Lu Wang

Abstract: For any asymptotically conical self-shrinker with entropy less than or equal to that of a cylinder we show that the link of the asymptotic cone must separate the unit sphere into exactly two connected components, both diffeomorphic to the self-shrinker. Combining this with recent work of Brendle, we conclude that the round sphere uniquely minimizes the entropy among all non-flat two-dimensional se… ▽ More For any asymptotically conical self-shrinker with entropy less than or equal to that of a cylinder we show that the link of the asymptotic cone must separate the unit sphere into exactly two connected components, both diffeomorphic to the self-shrinker. Combining this with recent work of Brendle, we conclude that the round sphere uniquely minimizes the entropy among all non-flat two-dimensional self-shrinkers. This confirms a conjecture of Colding-Ilmanen-Minicozzi-White in dimension two. △ Less

Submitted 28 June, 2016; v1 submitted 8 April, 2015; originally announced April 2015.

Comments: 25 pages, 1 figure. Added figure

MSC Class: 53C44

Journal ref: Duke Math. J. 166, no. 3 (2017), 403-435

Stacks in Representation Theory. What is a continuous representation of an algebraic group ?

Authors: Joseph Bernstein

Abstract: In this note I introduce a new approach to (or rather a new language for) representation theory of groups. Namely, I propose to consider a (complex) representation of a group $G$ as a sheaf on some geometric object (a stack). This point of view necessarily leads to a conclusion that the standard approach to (continuous) representations of algebraic groups is ideologically inconsistent. I propose… ▽ More In this note I introduce a new approach to (or rather a new language for) representation theory of groups. Namely, I propose to consider a (complex) representation of a group $G$ as a sheaf on some geometric object (a stack). This point of view necessarily leads to a conclusion that the standard approach to (continuous) representations of algebraic groups is ideologically inconsistent. I propose a way to modify this approach using sheaves on stacks. In new version I corrected some misprints and added an explanation how stack's approach is related to Vogan's picture of representations. △ Less

Submitted 12 May, 2016; v1 submitted 1 October, 2014; originally announced October 2014.

MSC Class: 18F; 14A; 20G; 20L

A Sharp Lower Bound for the Entropy of Closed Hypersurfaces up to Dimension Six

Authors: Jacob Bernstein, Lu Wang

Abstract: In [5], Colding-Ilmanen-Minicozzi-White showed that within the class of closed smooth self-shrinkers in $\mathbb{R}^{n+1}$, the entropy is uniquely minimized at the round sphere. They conjectured that, for $2\leq n\leq 6$, the round sphere minimizes the entropy among all closed smooth hypersurfaces. Using an appropriate weak mean curvature flow, we prove their conjecture. For these dimensions, our… ▽ More In [5], Colding-Ilmanen-Minicozzi-White showed that within the class of closed smooth self-shrinkers in $\mathbb{R}^{n+1}$, the entropy is uniquely minimized at the round sphere. They conjectured that, for $2\leq n\leq 6$, the round sphere minimizes the entropy among all closed smooth hypersurfaces. Using an appropriate weak mean curvature flow, we prove their conjecture. For these dimensions, our approach also gives a new proof of the main result of [5] and extends its conclusions to compact singular self-shrinkers. △ Less

Submitted 28 June, 2016; v1 submitted 11 June, 2014; originally announced June 2014.

Comments: 20 pages. Streamlined Section 4

MSC Class: 53C44

One-dimensional projective structures, convex curves and the ovals of Benguria & Loss

Authors: Jacob Bernstein, Thomas Mettler

Abstract: Benguria and Loss have conjectured that, amongst all smooth closed curves of length $2π$ in the plane, the lowest possible eigenvalue of the operator $L=-Δ+κ^2$ was one. They observed that this value was achieved on a two-parameter family, $\mathcal{O}$, of geometrically distinct ovals containing the round circle and collapsing to a multiplicity-two line segment. We characterize the curves in… ▽ More Benguria and Loss have conjectured that, amongst all smooth closed curves of length $2π$ in the plane, the lowest possible eigenvalue of the operator $L=-Δ+κ^2$ was one. They observed that this value was achieved on a two-parameter family, $\mathcal{O}$, of geometrically distinct ovals containing the round circle and collapsing to a multiplicity-two line segment. We characterize the curves in $\mathcal{O}$ as absolute minima of two related geometric functionals. We also discuss a connection with projective differential geometry and use it to explain the natural symmetries of all three problems. △ Less

Submitted 27 August, 2014; v1 submitted 31 March, 2014; originally announced March 2014.

Comments: 18 pages, exposition improved

Journal ref: Comm. Math. Phys. 336 (2015), no. 2, 933-952

A Remark on a Uniqueness Property of High Multiplicity Tangent Flows in Dimension Three

Authors: Jacob Bernstein, Lu Wang

Abstract: In this note, we combine the work of Ilmanen and of Colding-Ilmanen-Minicozzi to observe a uniqueness property for tangent flows at the first singular time of a smooth mean curvature flow of a closed surface in 3-dimensional Euclidean space. Specifically, if, at a fixed singular point, one tangent flow is a positive integer multiple of a shrinking plane, cylinder or sphere, then, modulo rotations,… ▽ More In this note, we combine the work of Ilmanen and of Colding-Ilmanen-Minicozzi to observe a uniqueness property for tangent flows at the first singular time of a smooth mean curvature flow of a closed surface in 3-dimensional Euclidean space. Specifically, if, at a fixed singular point, one tangent flow is a positive integer multiple of a shrinking plane, cylinder or sphere, then, modulo rotations, all tangent flows at the point are the same. △ Less

Submitted 26 February, 2014; originally announced February 2014.

Comments: 6 pages, 0 figure

MSC Class: 53C44 (Primary)

Periods and global invariants of automorphic representations

Authors: Joseph Bernstein, Andre Reznikov

Abstract: We consider periods of automorphic representations of adele groups defined by integrals along Gelfand subgroups. We define natural maps between local components of such periods and construct corresponding global maps using automorphic $L$-functions. This leads to an introduction of a global invariant of an automorphic representation arising from two such periods. We compute this invariant in some… ▽ More We consider periods of automorphic representations of adele groups defined by integrals along Gelfand subgroups. We define natural maps between local components of such periods and construct corresponding global maps using automorphic $L$-functions. This leads to an introduction of a global invariant of an automorphic representation arising from two such periods. We compute this invariant in some cases. △ Less

Submitted 2 October, 2022; v1 submitted 10 December, 2013; originally announced December 2013.

Comments: To appear

Journal ref: J. of Number Theory (2022)

Topological Type of Limit Laminations of Embedded Minimal Disks

Authors: Jacob Bernstein, Giuseppe Tinaglia

Abstract: We consider two natural classes of minimal laminations in three-manifolds. Both classes may be thought of as limits - in different senses - of embedded minimal disks. In both cases, we prove that, under a natural geometric assumption on the three-manifold, the leaves of these laminations are topologically either disks, annuli or Moebius bands. This answers a question posed by Hoffman and White. We consider two natural classes of minimal laminations in three-manifolds. Both classes may be thought of as limits - in different senses - of embedded minimal disks. In both cases, we prove that, under a natural geometric assumption on the three-manifold, the leaves of these laminations are topologically either disks, annuli or Moebius bands. This answers a question posed by Hoffman and White. △ Less

Submitted 26 May, 2016; v1 submitted 24 September, 2013; originally announced September 2013.

Comments: 21 pages. Published version

MSC Class: 53A10 (Primary)

Journal ref: J. Differential Geom. 102 (2016), no. 1, 1--23

Two-dimensional gradient Ricci solitons revisited

Authors: Jacob Bernstein, Thomas Mettler

Abstract: In this note, we complete the classification of the geometry of non-compact two-dimensional gradient Ricci solitons. As a consequence, we obtain two corollaries: First, a complete two-dimensional gradient Ricci soliton has bounded curvature. Second, we give examples of complete two-dimensional expanding Ricci solitons with negative curvature that are topologically disks and are not hyperbolic spac… ▽ More In this note, we complete the classification of the geometry of non-compact two-dimensional gradient Ricci solitons. As a consequence, we obtain two corollaries: First, a complete two-dimensional gradient Ricci soliton has bounded curvature. Second, we give examples of complete two-dimensional expanding Ricci solitons with negative curvature that are topologically disks and are not hyperbolic space. △ Less

Submitted 28 March, 2013; v1 submitted 27 March, 2013; originally announced March 2013.

Comments: 15 pages, 3 figures. Corrected statement and proof of Theorem 2 and removed Corollary 1

MSC Class: 53C44

Journal ref: Int. Math. Res. Not. 2015 (2015), no. 1, 78-98

Regarding a uniqueness property of singly-periodic Scherk surfaces

Authors: Jacob Bernstein

Abstract: Inspired by an argument of Ros [15] -- we use the López-Ros deformation to give another proof of the fact -- due to Meeks and Wolf [13] -- that the only smooth, connected, singly-periodic minimal surfaces in $\Real^3$ with the area growth of two planes are the singly-periodic Scherk surfaces. Inspired by an argument of Ros [15] -- we use the López-Ros deformation to give another proof of the fact -- due to Meeks and Wolf [13] -- that the only smooth, connected, singly-periodic minimal surfaces in $\Real^3$ with the area growth of two planes are the singly-periodic Scherk surfaces. △ Less

Submitted 13 May, 2013; v1 submitted 18 March, 2013; originally announced March 2013.

Comments: This paper has been withdrawn by the author due to the incorrect and overly strong statement of Proposition B.3. This leads to a serious gap in the proof of Proposition 3.5

MSC Class: 53A10

Characterizing classical minimal surfaces via the entropy differential

Authors: Jacob Bernstein, Thomas Mettler

Abstract: We introduce on any smooth oriented minimal surface in Euclidean $3$-space a meromorphic quadratic differential, $P$, which we call the entropy differential. This differential arises naturally in a number of different contexts. Of particular interest is the realization of its real part as a conservation law for a natural geometric functional -- which is, essentially, the entropy of the Gauss curva… ▽ More We introduce on any smooth oriented minimal surface in Euclidean $3$-space a meromorphic quadratic differential, $P$, which we call the entropy differential. This differential arises naturally in a number of different contexts. Of particular interest is the realization of its real part as a conservation law for a natural geometric functional -- which is, essentially, the entropy of the Gauss curvature. We characterize several classical surfaces -- including Enneper's surface, the catenoid and the helicoid -- in terms of $P$. As an application, we prove a novel curvature estimate for embedded minimal surfaces with small entropy differential and an associated compactness theorem. △ Less

Submitted 11 August, 2016; v1 submitted 8 January, 2013; originally announced January 2013.

Comments: 27 pages. Revised version

MSC Class: 53A10; 70S10

Journal ref: J. Geom. Anal. 27 (2017), no. 3, 2235-2268

Some Singular Limit Laminations of Embedded Minimal Planar Domains

Authors: Jacob Bernstein

Abstract: In this paper we give two examples of sequences of embedded minimal planar domains in $\mathbb{R}^3$ which converge to singular laminations of $\mathbb{R}^3$. In contrast with the situation for embedded minimal disks, these examples do not arise from complete embedded minimal planar domains and highlight some of the subtleties inherent in understanding refined properties of embedded minimal planar… ▽ More In this paper we give two examples of sequences of embedded minimal planar domains in $\mathbb{R}^3$ which converge to singular laminations of $\mathbb{R}^3$. In contrast with the situation for embedded minimal disks, these examples do not arise from complete embedded minimal planar domains and highlight some of the subtleties inherent in understanding refined properties of embedded minimal planar domains. △ Less

Submitted 19 July, 2011; originally announced July 2011.

Comments: 17 pages, 1 figure

MSC Class: 53A10

Journal ref: Bernstein, Jacob. Int. Math. Res. Not. 2012 (2012), no. 18, 4301-4324

A Variational Characterization of the Catenoid

Authors: Jacob Bernstein, Christine Breiner

Abstract: In this note, we use a result of Osserman and Schiffer \cite{OS} to give a variational characterization of the catenoid. Namely, we show that subsets of the catenoid minimize area within a geometrically natural class of minimal annuli. To the best of our knowledge, this fact has gone unremarked upon in the literature. As an application of the techniques, we give a sharp condition on the lengths of… ▽ More In this note, we use a result of Osserman and Schiffer \cite{OS} to give a variational characterization of the catenoid. Namely, we show that subsets of the catenoid minimize area within a geometrically natural class of minimal annuli. To the best of our knowledge, this fact has gone unremarked upon in the literature. As an application of the techniques, we give a sharp condition on the lengths of a pair of connected, simple closed curves $σ_1$ and $σ_2$ lying in parallel planes that precludes the existence of a connected minimal surface $Σ$ with $\partial Σ=σ_1\cupσ_2$. △ Less

Submitted 23 February, 2011; v1 submitted 17 December, 2010; originally announced December 2010.

Comments: 21 pages, 1 figure, added Section 4 with additional application

MSC Class: 53A10

Journal ref: Calc. Var. and PDE. 49 (2014), no. 1-2, 215-232

Symmetry of Embedded Genus-One Helicoids

Authors: Jacob Bernstein, Christine Breiner

Abstract: In this note, we use the Lopez-Ros deformation introduced in [9] to show that any embedded genus-one helicoid must be symmetric with respect to rotation by 180 degrees around a normal line. This partially answers a conjecture of Bobenko from [3]. We also show this symmetry holds for an embedded genus-k helicoid $Σ$, provided the underlying conformal structure of $Σ$ is hyperelliptic. In this note, we use the Lopez-Ros deformation introduced in [9] to show that any embedded genus-one helicoid must be symmetric with respect to rotation by 180 degrees around a normal line. This partially answers a conjecture of Bobenko from [3]. We also show this symmetry holds for an embedded genus-k helicoid $Σ$, provided the underlying conformal structure of $Σ$ is hyperelliptic. △ Less

Submitted 16 November, 2010; originally announced November 2010.

Comments: 11 pages, 1 figure

MSC Class: 53A10; 49Q05

Journal ref: Duke Math. J. 159, no. 1 (2011), 83-97