BZPEERSource author record
Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.
Catalog footprint
Published work
In this paper, we prove that the Euclidean distance between two independent random vectors uniformly distributed on lpn-balls (1≤p≤∞) or on its boundary satisfies a central limit theorem as n tends to ∞. Also, we give a compact proof of the case of the sphere, which was proved by Hammersley. Furthermore, we complement our central limit theorem by providing large deviation principles for the cases p≥2.
In this article we take a probabilistic look at Hölder's inequality, considering the ratio of terms in the classical Hölder inequality for random vectors in Rn. We prove a central limit theorem for this ratio, which then allows us to reverse the inequality up to a multiplicative constant with high probability. The models of randomness include the uniform distribution on ℓpn balls and spheres. We also provide a Berry-Esseen type result and prove a large and a moderate deviation principle for the suitably normalized Hölder ratio.
We prove upper and lower bounds on the minimal spherical dispersion, improving upon previous estimates obtained by Rote and Tichy [Spherical dispersion with an application to polygonal approximation of curves, Anz. Österreich. Akad. Wiss. Math.-Natur. Kl. 132 (1995), 3--10]. In particular, we see that the inverse N(ε,d) of the minimal spherical dispersion is, for fixed ε>0, linear in the dimension d of the ambient space. We also derive upper and lower bounds on the expected dispersion for points chosen independently and uniformly at random from the Euclidean unit sphere. In terms of the corresponding inverse N(ε,d), our bounds are optimal with respect to the dependence on ε.
Let 0<p,q≤∞ and denote by SpN and SqN the corresponding Schatten classes of real N×N matrices. We study approximation quantities of natural identities SpN↪SqN between Schatten classes and prove asymptotically sharp bounds up to constants only depending on p and q, showing how approximation numbers are intimately related to the Gelfand numbers and their duals, the Kolmogorov numbers. In particular, we obtain new bounds for those sequences of s-numbers. Our results improve and complement bounds previously obtained by B. Carl and A. Defant [J. Approx. Theory, 88(2):228--256, 1997], Y. Gordon, H. König, and C. Schütt [J. Approx. Theory, 49(3):219--239, 1987], A. Hinrichs and C. Michels [Rend. Circ. Mat. Palermo (2) Suppl., (76):395--411, 2005], and A. Hinrichs, J. Prochno, and J. Vybíral [preprint, 2020]. We also treat the case of quasi-Schatten norms, which is relevant in applications such as low-rank matrix recovery.
We study the asymptotics of the k-regular self-similar fragmentation process. For α>0 and an integer k≥2, this is the Markov process (It)t≥0 in which each It is a union of open subsets of [0,1), and independently each subinterval of It of size u breaks into k equally sized pieces at rate uα. Let k−mt and k−Mt be the respective sizes of the largest and smallest fragments in It. By relating (It)t≥0 to a branching random walk, we find that there exist explicit deterministic functions g(t) and h(t) such that ∣mt−g(t)∣≤1 and ∣Mt−h(t)∣≤1 for all sufficiently large t. Furthermore, for each n, we study the final time at which fragments of size k−n exist. In particular, by relating our branching random walk to a certain point process, we show that, after suitable rescaling, the laws of these times converge to a Gumbel distribution as n→∞.
Having its origin in theoretical computer science, the Kannan-Lovász-Simonovits (KLS) conjecture is one of the major open problems in asymptotic convex geometry and high-dimensional probability theory today. In this work, we establish a new connection between this conjecture and the study of large and moderate deviations for isotropic log-concave random vectors, thereby providing a novel possibility to tackle the conjecture. We then study the moderate deviations for the Euclidean norm of random orthogonally projected random vectors in an ℓpn-ball. This leads to a number of interesting observations: (A) the ℓ1n-ball is critical for the new approach; (B) for p≥2 the rate function in the moderate deviations principle undergoes a phase transition, depending on whether the scaling is below the square-root of the subspace dimensions or comparable; (C) for 1≤p<2 and comparable subspace dimensions, the rate function again displays a phase transition depending on its growth relative to np/2.
In this work the ℓq-norms of points chosen uniformly at random in a centered regular simplex in high dimensions are studied. Berry-Esseen bounds in the regime 1≤q<∞ are derived and complemented by a non-central limit theorem together with moderate and large deviations in the case where q=∞. A comparison with corresponding results for ℓpn-balls is carried out as well.
We study the precise asymptotic volume of balls in Orlicz spaces and show that the volume of the intersection of two Orlicz balls undergoes a phase transition when the dimension of the ambient space tends to infinity. This generalizes a result of Schechtman and Schmuckenschläger [GAFA, Lecture notes in Math. 1469 (1991), 174--178] for ℓpd-balls. As another application, we determine the precise asymptotic volume ratio for 2-concave Orlicz spaces ℓMd. Our method rests on ideas from statistical mechanics and large deviations theory, more precisely the maximum entropy or Gibbs principle for non-interacting particles, and presents a natural approach and fresh perspective to such geometric and volumetric questions. In particular, our approach explains how the p-generalized Gaussian distribution occurs in problems related to the geometry of ℓpd-balls, which are Orlicz balls when the Orlicz function is M(t)=∣t∣p.
We show that the minimal discrepancy of a point set in the d-dimensional unit cube with respect to Orlicz norms can exhibit both polynomial and weak tractability. In particular, we show that the ψα-norms of exponential Orlicz spaces are polynomially tractable.
Irrational numbers are introduced usually already introduced in lower secondary level schools. But typically, maybe with the exception of 2, no mathematical proof of irrationality is provided. In particular it is not proven that famous Euler's number e as well as the number π are irrational. In this article we want to show how this can be done with very elementary methods from calculus. In addition, we offer geometrical variants for many of the analytical statements, which in particular create variability in the level of requirements. ----- Irrationale Zahlen werden in der Schule bereits in der Sekundarstufe I eingeführt. Allerdings wird typischerweise, mit Ausnahme vielleicht für 2, kein mathematischer Beweis zur Irrationalität geführt. Insbesondere wird nicht bewiesen, dass die berühmte Eulersche Zahl e sowie die Kreiszahl π irrationale Zahlen sind. In diesem Artikel wollen wir aufzeigen, wie dies mit recht elementaren Methoden der Analysis möglich ist. Darüber hinaus bieten wir für viele der analytischen Aussagen geometrische Varianten zur Veranschaulichung, die insbesondere Variabilität im Anspruchsniveau schaffen.
Let 0<p,q≤∞ and denote by SpN and SqN the corresponding finite-dimensional Schatten classes. We prove optimal bounds, up to constants only depending on p and q, for the entropy numbers of natural embeddings between SpN and SqN. This complements the known results in the classical setting of natural embeddings between finite-dimensional ℓp spaces due to Schütt, Edmunds-Triebel, Triebel and Guédon-Litvak/Kühn. We present a rather short proof that uses all the known techniques as well as a constructive proof of the upper bound in the range N≤n≤N2 that allows deeper structural insight and is therefore interesting in its own right. Our main result can also be used to provide an alternative proof of recent lower bounds in the area of low-rank matrix recovery.
We prove estimates for the expected value of operator norms of Gaussian random matrices with independent and mean-zero entries, acting as operators from ℓp∗m to ℓqn, 1≤p∗≤2≤q≤∞.
In this work, we study the volume ratio of the projective tensor products ℓpn⊗πℓqn⊗πℓrn with 1≤p≤q≤r≤∞. We obtain asymptotic formulas that are sharp in almost all cases. As a consequence of our estimates, these spaces allow for a nearly Euclidean decomposition of Kashin type whenever 1≤p≤q≤r≤2 or 1≤p≤2≤r≤∞ and q=2. Also, from the Bourgain-Milman bound on the volume ratio of Banach spaces in terms of their cotype 2 constant, we obtain information on the cotype of these 3-fold projective tensor products. Our results naturally generalize to k-fold products ℓp1n⊗π⋯⊗πℓpkn with k∈N and 1≤p1≤⋯≤pk≤∞.
We prove some probabilistic estimates for tensor products of random vectors. As an application we obtain embeddings of certain matrix spaces into L1.
We prove some "high probability" results on the expected value of the mean width for random perturbations of random polytopes. The random perturbations are considered for Gaussian and p-stable random vectors, as well as uniform distributions on ℓpN-balls and the unit sphere.
In this note we prove that n!1∑π(∑i=1n∣xiai,π(i)∣2)1/2 is equivalent to a Musielak-Orlicz norm $\norm{x}_{\sum M_i}$. We also obtain the inverse result, i.e., given the Orlicz functions, we provide a formula for the choice of the matrix that generates the corresponding Musielak-Orlicz norm. As a consequence, we obtain the embedding of 2-concave Musielak-Orlicz spaces into L_1.
Given a normalized Orlicz function M we provide an easy formula for a distribution such that, if X is a random variable distributed accordingly and X1,...,Xn are independent copies of X, then the expected value of the p-norm of the vector (xiXi)i=1n is of the order ∥x∥M (up to constants dependent on p only). In case p=2 we need the function t↦tM′(t)−M(t) to be 2-concave and as an application immediately obtain an embedding of the corresponding Orlicz spaces into L1[0,1]. We also provide a general result replacing the ℓp-norm by an arbitrary N-norm. This complements some deep results obtained by Gordon, Litvak, Schütt, and Werner. We also prove a result in the spirit of their work which is of a simpler form and easier to apply. All results are true in the more general setting of Musielak-Orlicz spaces.
In this paper we show that, using combinatorial inequalities and Matrix-Averages, we can generate Musielak-Orlicz spaces, i.e., we prove that $1/π\sum_π \max\limits_{1 \leq i \leq n} \abs{x_i y_{iπ(i)}} \sim \norm{x}_{ΣM_i}$, where the Orlicz functions M1,...,Mn depend on the matrix (yij)i,j=1n. We also provide an approximation result for Musielak-Orlicz norms which already in the case of Orlicz spaces turned out to be very useful.
Given an Orlicz function M, we show which random variables ξi, i=1,...,n generate the associated Orlicz norm, i.e., which random variables yield $\mathbb{E} \max\limits_{1\leq i \leq n}|x_iξ_i| \sim \norm{(x_i)_{i=1}^n}_M$. As a corollary we obtain a representation for the distribution function in terms of M and M′ which can be easily applied to many examples of interest.
Let M and N be Orlicz functions. We establish some combinatorial inequalities and show that the product spaces l^n_M(l^n_N) are uniformly isomorphic to subspaces of L_1 if M and N are "separated" by a function t^r, 1<r<2.
We study the expected value of support functions of random polytopes in a certain direction, where the random polytope is given by independent random vectors uniformly distributed in an isotropic convex body. All results are obtained by an utterly novel approach, using probabilistic estimates in connection with Orlicz norms that were not used in this connection before.
In this paper we introduce a new sequence of quantities for random polytopes. Let $K_N=\conv\{X_1,...,X_N\}$ be a random polytope generated by independent random vectors uniformly distributed in an isotropic convex body K of Rn. We prove that the so-called k-th mean outer radius Rk(KN) has order max{k,logN}LK with high probability if n2≤N≤en. We also show that this is also the right order of the expected value of Rk(KN) in the full range n≤N≤en.
We show that the expected value of the mean width of a random polytope generated by N random vectors (n≤N≤en) uniformly distributed in an isotropic convex body in Rn is of the order logNLK. This completes a result of Dafnis, Giannopoulos and Tsolomitis. We also prove some results in connection with the 1-dimensional marginals of the uniform probability measure on an isotropic convex body, extending the interval in which the average of the distribution functions of those marginals behaves in a sub- or supergaussian way.
Let ξi, i=1,...,n, and ηj, j=1,...,m be iid p-stable respectively q-stable random variables, 1<p<q<2. We prove estimates for $\Ex_{Ω_1} \Ex_{Ω_2}\max_{i,j}\abs{a_{ij}ξ_i(ω_1)η_j(ω_2)}$ in terms of the ℓpm(ℓqn)-norm of (aij)i,j. Additionally, for p-stable and standard gaussian random variables we prove estimates in terms of the ℓpm(ℓMξn)-norm, Mξ depending on the Gaussians. Furthermore, we show that a sequence ξi, i=1,...,n of iid log−γ(1,p) distributed random variables (p≥2) generates a truncated ℓp-norm, especially $\Ex \max_{i}\abs{a_iξ_i}\sim \norm{(a_i)_i}_2$ for p=2. As far as we know, the generating distribution for ℓp-norms with p≥2 has not been known up to now.
Research graph
Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.
BZPEER is loading the nearby papers, people, topics and institutions for this page.