Showing 1–12 of 12 results for author: Vu, V H

Heterogeneous Hypergraph Embedding for Recommendation Systems

Authors: Darnbi Sakong, Viet Hung Vu, Thanh Trung Huynh, Phi Le Nguyen, Hongzhi Yin, Quoc Viet Hung Nguyen, Thanh Tam Nguyen

Abstract: Recent advancements in recommender systems have focused on integrating knowledge graphs (KGs) to leverage their auxiliary information. The core idea of KG-enhanced recommenders is to incorporate rich semantic information for more accurate recommendations. However, two main challenges persist: i) Neglecting complex higher-order interactions in the KG-based user-item network, potentially leading to… ▽ More Recent advancements in recommender systems have focused on integrating knowledge graphs (KGs) to leverage their auxiliary information. The core idea of KG-enhanced recommenders is to incorporate rich semantic information for more accurate recommendations. However, two main challenges persist: i) Neglecting complex higher-order interactions in the KG-based user-item network, potentially leading to sub-optimal recommendations, and ii) Dealing with the heterogeneous modalities of input sources, such as user-item bipartite graphs and KGs, which may introduce noise and inaccuracies. To address these issues, we present a novel Knowledge-enhanced Heterogeneous Hypergraph Recommender System (KHGRec). KHGRec captures group-wise characteristics of both the interaction network and the KG, modeling complex connections in the KG. Using a collaborative knowledge heterogeneous hypergraph (CKHG), it employs two hypergraph encoders to model group-wise interdependencies and ensure explainability. Additionally, it fuses signals from the input graphs with cross-view self-supervised learning and attention mechanisms. Extensive experiments on four real-world datasets show our model's superiority over various state-of-the-art baselines, with an average 5.18\% relative improvement. Additional tests on noise resilience, missing data, and cold-start problems demonstrate the robustness of our KHGRec framework. Our model and evaluation datasets are publicly available at \url{https://github.com/viethungvu1998/KHGRec}. △ Less

Submitted 4 July, 2024; originally announced July 2024.

Covariance Fitting Interferometric Phase Linking: Modular Framework and Optimization Algorithms

Authors: Phan Viet Hoa Vu, Arnaud Breloy, Frédéric Brigui, Yajing Yan, Guillaume Ginolhac

Abstract: Interferometric phase linking (IPL) has become a prominent technique for processing images of areas containing distributed scaterrers in SAR interferometry. Traditionally, IPL consists in estimating consistent phase differences between all pairs of SAR images in a time series from the sample covariance matrix of pixel patches on a sliding window. This paper reformulates this task as a covariance f… ▽ More Interferometric phase linking (IPL) has become a prominent technique for processing images of areas containing distributed scaterrers in SAR interferometry. Traditionally, IPL consists in estimating consistent phase differences between all pairs of SAR images in a time series from the sample covariance matrix of pixel patches on a sliding window. This paper reformulates this task as a covariance fitting problem: in this setup, IPL appears as a form of projection of an input covariance matrix so that it satisfies the phase closure property. Given this modular formulation, we propose an overview of covariance matrix estimates, regularization options, and matrix distances, that can be of interest when processing multi-temporal SAR data. In particular, we will observe that most of the existing IPL algorithms appear as special instances of this framework. We then present tools to efficiently solve related optimization problems on the torus of phase-only complex vectors: majorization-minimization and Riemannian optimization. We conclude by illustrating the merits of different options on a real-world case study. △ Less

Submitted 13 March, 2024; originally announced March 2024.

Normal vector of a random hyperplane

Authors: Hoi H. Nguyen, Van H. Vu

Abstract: Let v_1,...,v_{n-1} be n-1 independent vectors in R^n (or C^n). We study x, the unit normal vector of the hyperplane spanned by the v_i. Our main finding is that x resembles a random vector chosen uniformly from the unit sphere, under some randomness assumption on the v_i. Our result has applications in random matrix theory. Consider an n by n random matrix with iid entries. We first prove an ex… ▽ More Let v_1,...,v_{n-1} be n-1 independent vectors in R^n (or C^n). We study x, the unit normal vector of the hyperplane spanned by the v_i. Our main finding is that x resembles a random vector chosen uniformly from the unit sphere, under some randomness assumption on the v_i. Our result has applications in random matrix theory. Consider an n by n random matrix with iid entries. We first prove an exponential bound on the upper tail for the least singular value, improving the earlier linear bound by Rudelson and Vershynin. Next, we derive optimal delocalization for the eigenvectors corresponding to eigenvalues of small modulus. △ Less

Submitted 17 April, 2016; originally announced April 2016.

Comments: 24 pages, 2 figures. A portion of this work substantially improves a previous preprint of the first author

Non-abelian Littlewood-Offord inequalities

Authors: Pham H. Tiep, Van H. Vu

Abstract: In 1943, Littlewood and Offord proved the first anti-concentration result for sums of independent random variables. Their result has since then been strengthened and generalized by generations of researchers, with applications in several areas of mathematics. In this paper, we present the first non-abelian analogue of Littlewood-Offord result, a sharp anti-concentration inequality for products o… ▽ More In 1943, Littlewood and Offord proved the first anti-concentration result for sums of independent random variables. Their result has since then been strengthened and generalized by generations of researchers, with applications in several areas of mathematics. In this paper, we present the first non-abelian analogue of Littlewood-Offord result, a sharp anti-concentration inequality for products of independent random variables. △ Less

Submitted 6 August, 2015; v1 submitted 5 June, 2015; originally announced June 2015.

Comments: 14 pages Second version. Dependence of the upper bound on the matrix size in the main results has been removed

MSC Class: 05D40; 20C33 (Primary); 60B15; 60C05; 60G50 (Secondary)

Small ball probability, Inverse theorems, and applications

Authors: Hoi H. Nguyen, Van H. Vu

Abstract: Let $ξくしー$ be a real random variable with mean zero and variance one and $A={a_1,...,a_n}$ be a multi-set in $\R^d$. The random sum $$S_A := a_1 ξくしー_1 + ... + a_n ξくしー_n $$ where $ξくしー_i$ are iid copies of $ξくしー$ is of fundamental importance in probability and its applications. We discuss the small ball problem, the aim of which is to estimate the maximum probability that $S_A$ belongs to a ball with given… ▽ More Let $ξくしー$ be a real random variable with mean zero and variance one and $A={a_1,...,a_n}$ be a multi-set in $\R^d$. The random sum $$S_A := a_1 ξくしー_1 + ... + a_n ξくしー_n $$ where $ξくしー_i$ are iid copies of $ξくしー$ is of fundamental importance in probability and its applications. We discuss the small ball problem, the aim of which is to estimate the maximum probability that $S_A$ belongs to a ball with given small radius, following the discovery made by Littlewood-Offord and Erdos almost 70 years ago. We will mainly focus on recent developments that characterize the structure of those sets $A$ where the small ball probability is relatively large. Applications of these results include full solutions or significant progresses of many open problems in different areas. △ Less

Submitted 31 December, 2012; originally announced January 2013.

Comments: 47 pages

Mapping Incidences

Authors: Van H. Vu, Melanie Matchett Wood, Philip Matchett Wood

Abstract: We show that any finite set S in a characteristic zero integral domain can be mapped to the finite field of order p, for infinitely many primes p, preserving all algebraic incidences in S. This can be seen as a generalization of the well-known Freiman isomorphism lemma, and we give several combinatorial applications (such as sum-product estimates). We show that any finite set S in a characteristic zero integral domain can be mapped to the finite field of order p, for infinitely many primes p, preserving all algebraic incidences in S. This can be seen as a generalization of the well-known Freiman isomorphism lemma, and we give several combinatorial applications (such as sum-product estimates). △ Less

Submitted 15 April, 2011; v1 submitted 27 November, 2007; originally announced November 2007.

Comments: 15 pages, to appear in the Journal of the London Mathematical Society. Section 3 on Erdős distance problem from the previous version has been removed, since the most current version of arXiv:math/0301343v3 [math.CO] has the added restriction that -1 is not a square. Other minor revisions were also made

MSC Class: 05B25; 11T99

Subset sums in $\BZ_p$

Authors: H. H. Nguyen, E. Szemeredi, V. H. Vu

Abstract: Let $\BZ_p$ be the finite field of prime order $p$ and $A$ be a subset of $\BZ_p$. We prove several sharp results about the following two basic questions: (1) When can one represent zero as a sum of distinct elements of $A$ ? (2) When can one represent every element of $\BZ_p$ as a sum of distinct elements of $A$ ? Let $\BZ_p$ be the finite field of prime order $p$ and $A$ be a subset of $\BZ_p$. We prove several sharp results about the following two basic questions: (1) When can one represent zero as a sum of distinct elements of $A$ ? (2) When can one represent every element of $\BZ_p$ as a sum of distinct elements of $A$ ? △ Less

Submitted 5 October, 2006; originally announced October 2006.

Central limit theorems for Gaussian polytopes

Authors: I. Barany, V. H. Vu

Abstract: Choose $n$ random, independent points in $\R^d$ according to the standard normal distribution. Their convex hull $K_n$ is the {\sl Gaussian random polytope}. We prove that the volume and the number of faces of $K_n$ satisfy the central limit theorem, settling a well known conjecture in the field. Choose $n$ random, independent points in $\R^d$ according to the standard normal distribution. Their convex hull $K_n$ is the {\sl Gaussian random polytope}. We prove that the volume and the number of faces of $K_n$ satisfy the central limit theorem, settling a well known conjecture in the field. △ Less

Submitted 5 October, 2006; originally announced October 2006.

Comments: to appear in Annals of Probability

Structure of large incomplete sets in abelian groups

Authors: Van H. Vu

Abstract: Let $G$ be a finite abelian group and $A$ be a subset of $G$. We say that $A$ is complete if every element of $G$ can be represented as a sum of different elements of $A$. In this paper, we study the following question: {\it What is the structure of a large incomplete set ?} The typical answer is that such a set is essentially contained in a maximal subgroup. As a by-product, we obtain a new pro… ▽ More Let $G$ be a finite abelian group and $A$ be a subset of $G$. We say that $A$ is complete if every element of $G$ can be represented as a sum of different elements of $A$. In this paper, we study the following question: {\it What is the structure of a large incomplete set ?} The typical answer is that such a set is essentially contained in a maximal subgroup. As a by-product, we obtain a new proof for several earlier results. △ Less

Submitted 4 October, 2006; originally announced October 2006.

The Rank of Random Graphs

Authors: Kevin P. Costello, Van H. Vu

Abstract: We show that almost surely the rank of the adjacency matrix of the Erdös-Rényi random graph $G(n,p)$ equals the number of non-isolated vertices for any $c\ln n/n<p<1/2$, where $c$ is an arbitrary positive constant larger than 1/2. In particular, the giant component (a.s.) has full rank in this range. We show that almost surely the rank of the adjacency matrix of the Erdös-Rényi random graph $G(n,p)$ equals the number of non-isolated vertices for any $c\ln n/n<p<1/2$, where $c$ is an arbitrary positive constant larger than 1/2. In particular, the giant component (a.s.) has full rank in this range. △ Less

Submitted 16 June, 2006; originally announced June 2006.

Comments: 19 pages, no figures

MSC Class: 15A52

On the concentration of eigenvalues of random symmetric matrices

Authors: Michael Krivelevich, Van H. Vu

Abstract: We prove that few largest (and most important) eigenvalues of random symmetric matrices of various kinds are very strongly concentrated. This strong concentration enables us to compute the means of these eigenvalues with high precision. Our approach uses Talagrand's inequality and is very different from standard approaches. We prove that few largest (and most important) eigenvalues of random symmetric matrices of various kinds are very strongly concentrated. This strong concentration enables us to compute the means of these eigenvalues with high precision. Our approach uses Talagrand's inequality and is very different from standard approaches. △ Less

Submitted 21 September, 2000; originally announced September 2000.

MSC Class: 15A52 (Primary); 60B**; 05C80 (Secondary)

The cover time, the blanket time, and the Matthews bound

Authors: J. Kahn, J. H. Kim, L. Lovasz, V. H. Vu

Abstract: The cover time C of a graph G is the expected time for a random walk starting from the worst vertex to cover all vertices in G. Similarly, the blanket time B is the expected time to visit all vertices within a constant factor of number of times suggested by the stationary distribution. (Our definition will be slightly stronger than this.) Obviously, all vertices are covered when the graph is bla… ▽ More The cover time C of a graph G is the expected time for a random walk starting from the worst vertex to cover all vertices in G. Similarly, the blanket time B is the expected time to visit all vertices within a constant factor of number of times suggested by the stationary distribution. (Our definition will be slightly stronger than this.) Obviously, all vertices are covered when the graph is blanketed, and hence C <= B. The blanket time is introduced by Winkler and Zuckerman motivated by applications in Markov estimation and distributed computing. They conjectured B =O(C) and proved B=O(C ln n ). In this paper, we introduce another parameter M motivated by Matthews' theorem and prove M/2 <= C <= B = O(( M ln ln n)^2). In particular, B = O(C (ln ln n)^2). The lower bound is still valid for the cover time C(πぱい) starting from the stationary distribution. We also show that there is a polynomial time algorithm to approximate M within a factor of 2 and so does for C within a factor of O((ln ln n)^2), improving previous bound of O(ln n) of Matthews'. △ Less

Submitted 11 May, 2000; originally announced May 2000.

Comments: 9 page article on random walks