arXiv subjects

Minh Tang

Publications and source records attributed to Minh Tang.

At least 19 recordsLinked to original sources

arXiv2026

Predictive Subsampling for Scalable Inference in Networks

Current methods for statistical inference in networks often encounter substantial computational bottlenecks when applied to the massive network datasets that are increasingly common across scientific domains. In this paper, we develop \textit{Predictive Subsampling} (\texttt{PredSub}), a scalable framework for estimation and two-sample testing in networks. The central idea is to replace a full-sample estimation procedure by estimation on a random subsample, followed by out-of-sample prediction of the remaining vertices. This construction exploits the fact that the subsample provides an inferential anchor, which enables each remaining vertex to be incorporated in a \textit{predictive} manner through a fast vector operation. Building on this estimator, we develop two scalable procedures for two-sample testing, namely \texttt{PredSubTest} and \texttt{PureSubTest}. We establish finite-sample error bounds as well as estimation and testing consistency of the proposed methods in both Frobenius and two-to-infinity norms. These results formally characterize the trade-offs between statistical accuracy and computational efficiency with respect to the subsample size, the choice of test statistic, and the choice of norm. We demonstrate the empirical performance of the proposed methods through detailed simulation studies and two real-world applications involving DBLP coauthorship networks and the Cannes 2013 social media networks.

stat.ME↗

arXiv2025

A Unified Framework for Community Detection and Model Selection in Blockmodels

Blockmodels are a foundational tool for modeling community structure in networks, with the stochastic blockmodel (SBM), degree-corrected blockmodel (DCBM), and popularity-adjusted blockmodel (PABM) forming a natural hierarchy of increasing generality. While community detection under these models has been extensively studied, much less attention has been paid to the model selection problem, i.e., determining which model best fits a given network. Building on recent theoretical insights about the spectral geometry of these models, we propose a unified framework for simultaneous community detection and model selection across the full blockmodel hierarchy. A key innovation is the use of loss functions that serve a dual role: they act as objective functions for community detection and as test statistics for hypothesis testing. We develop a greedy algorithm to minimize these loss functions and establish theoretical guarantees for exact label recovery and model selection consistency under each model. Extensive simulation studies demonstrate that our method achieves high accuracy in both tasks, outperforming or matching state-of-the-art alternatives. Applications to five real-world networks further illustrate the interpretability and practical utility of our approach. R code for implementing the method is available at https://github.com/subhankarbhadra/model-selection.

stat.ME↗

arXiv2025

Out-of-Sample Embedding with Proximity Data: Projection versus Restricted Reconstruction

The problem of using proximity (similarity or dissimilarity) data for the purpose of "adding a point to a vector diagram" was first studied by J.C. Gower in 1968. Since then, a number of methods -- mostly kernel methods -- have been proposed for solving what has come to be called the problem of *out-of-sample embedding*. We survey the various kernel methods that we have encountered and show that each can be derived from one or the other of two competing strategies: *projection* or *restricted reconstruction*. Projection can be analogized to a well-known formula for adding a point to a principal component analysis. Restricted reconstruction poses a different challenge: how to best approximate redoing the entire multivariate analysis while holding fixed the vector diagram that was previously obtained. This strategy results in a nonlinear optimization problem that can be simplified to a unidimensional search. Various circumstances may warrant either projection or restricted reconstruction.

stat.ML↗

arXiv2025

Eigenvector fluctuations and limit results for random graphs with infinite rank kernels

This paper systematically studies the behavior of the leading eigenvectors for independent edge undirected random graphs generated from a general latent position model whose link function is possibly infinite rank and also possibly indefinite. We first derive uniform error bounds in the two-to-infinity norm as well as row-wise normal approximations for the leading sample eigenvectors. We then build on these results to tackle two graph inference problems, namely (i) entrywise bounds for graphon estimation and (ii) testing for the equality of latent positions, the latter of which is achieved by proposing a rank-adaptive test statistic that converges in distribution to a weighted sum of independent chi-square random variables under the null hypothesis. Our fine-grained theoretical guarantees and applications differ from the existing literature which primarily considers first order upper bounds and more restrictive low rank or positive semidefinite model assumptions. Further, our results collectively quantify the statistical properties of eigenvector-based spectral embeddings with growing dimensionality for large graphs.

math.ST↗

arXiv2024

Chain-linked multiple matrix integration via embedding alignment

Motivated by the increasing demand for multi-source data integration in various scientific fields, in this paper we study matrix completion in scenarios where the data exhibits certain block-wise missing structures -- specifically, where only a few noisy submatrices representing (overlapping) parts of the full matrix are available. We propose the Chain-linked Multiple Matrix Integration (CMMI) procedure to efficiently combine the information that can be extracted from these individual noisy submatrices. CMMI begins by deriving entity embeddings for each observed submatrix, then aligns these embeddings using overlapping entities between pairs of submatrices, and finally aggregates them to reconstruct the entire matrix of interest. We establish, under mild regularity conditions, entrywise error bounds and normal approximations for the CMMI estimates. Simulation studies and real data applications show that CMMI is computationally efficient and effective in recovering the full matrix, even when overlaps between the observed submatrices are minimal.

stat.ME↗

arXiv2024

Regression for matrix-valued data via Kronecker products factorization

We study the matrix-variate regression problem $Y_i = \sum_{k} \beta_{1k} X_i \beta_{2k}^{\top} + E_i$ for $i=1,2\dots,n$ in the high dimensional regime wherein the response $Y_i$ are matrices whose dimensions $p_{1}\times p_{2}$ outgrow both the sample size $n$ and the dimensions $q_{1}\times q_{2}$ of the predictor variables $X_i$ i.e., $q_{1},q_{2} \ll n \ll p_{1},p_{2}$. We propose an estimation algorithm, termed KRO-PRO-FAC, for estimating the parameters $\{\beta_{1k}\} \subset \Re^{p_1 \times q_1}$ and $\{\beta_{2k}\} \subset \Re^{p_2 \times q_2}$ that utilizes the Kronecker product factorization and rearrangement operations from Van Loan and Pitsianis (1993). The KRO-PRO-FAC algorithm is computationally efficient as it does not require estimating the covariance between the entries of the $\{Y_i\}$. We establish perturbation bounds between $\hat{\beta}_{1k} -\beta_{1k}$ and $\hat{\beta}_{2k} - \beta_{2k}$ in spectral norm for the setting where either the rows of $E_i$ or the columns of $E_i$ are independent sub-Gaussian random vectors. Numerical studies on simulated and real data indicate that our procedure is competitive, in terms of both estimation error and predictive accuracy, compared to other existing methods.

stat.ML↗

arXiv2023

Independence testing for inhomogeneous random graphs

Testing for independence between graphs is a problem that arises naturally in social network analysis and neuroscience. In this paper, we address independence testing for inhomogeneous Erd\H{o}s-R\'{e}nyi random graphs on the same vertex set. We first formulate a notion of pairwise correlations between the edges of these graphs and derive a necessary condition for their detectability. We next show that the problem can exhibit a statistical vs. computational tradeoff, i.e., there are regimes for which the correlations are statistically detectable but may require algorithms whose running time is exponential in n, the number of vertices. Finally, we consider a special case of correlation testing when the graphs are sampled from a latent space model (graphon) and propose an asymptotically valid and consistent test procedure that also runs in time polynomial in n.

stat.ME↗

arXiv2022

Adversarial contamination of networks in the setting of vertex nomination: a new trimming method

As graph data becomes more ubiquitous, the need for robust inferential graph algorithms to operate in these complex data domains is crucial. In many cases of interest, inference is further complicated by the presence of adversarial data contamination. The effect of the adversary is frequently to change the data distribution in ways that negatively affect statistical and algorithmic performance. We study this phenomenon in the context of vertex nomination, a semi-supervised information retrieval task for network data. Here, a common suite of methods relies on spectral graph embeddings, which have been shown to provide both good algorithmic performance and flexible settings in which regularization techniques can be implemented to help mitigate the effect of an adversary. Many current regularization methods rely on direct network trimming to effectively excise the adversarial contamination, although this direct trimming often gives rise to complicated dependency structures in the resulting graph. We propose a new trimming method that operates in model space which can address both block structure contamination and white noise contamination (contamination whose distribution is unknown). This model trimming is more amenable to theoretical analysis while also demonstrating superior performance in a number of simulations, compared to direct trimming.

stat.ML↗

arXiv2022

Limit results for distributed estimation of invariant subspaces in multiple networks inference and PCA

Several statistical problems, such as multiple heterogeneous graph analysis, distributed PCA, integrative data analysis, and simultaneous dimension reduction of images, can involve a collection of $m$ matrices whose leading subspaces $U^{(i)}$ consist of a shared subspace $U_c$ and individual subspaces $U_s^{(i)}$. We consider a distributed estimation procedure that first obtains $\hat U^{(i)}$ as the leading singular vectors for each observed noisy matrix, then computes the leading left singular vectors of the concatenated matrix $[\hat U^{(1)}|\hat U^{(2)}|\dots|\hat U^{(m)}]$ as $\hat U_c$, and finally computes the leading singular vectors of the projection of each $\hat U^{(i)}$ onto the orthogonal complement of $\hat U_c$ as $\hat U_s^{(i)}$. In this paper, we provide a framework for deriving limit results for such distributed estimation procedures, including expansions of estimation errors in both common and individual subspaces and their asymptotically normal approximations. We apply this framework specifically to (1) parameter estimation for multiple heterogeneous random graphs with shared subspaces, and (2) distributed PCA for independent sub-Gaussian random vectors with spiked covariance structures. Leveraging these results, we also consider a two-sample test for the null hypothesis that a pair of random graphs have the same edge probabilities, and present a test statistic whose limiting distribution converges to a central (resp., non-central) $\chi^2$ distribution under the null (resp., local alternative) hypothesis.

math.ST↗

arXiv2022

Perturbation Analysis of Randomized SVD and its Applications to Statistics

Randomized singular value decomposition (RSVD) is a class of computationally efficient algorithms for computing the truncated SVD of large data matrices. Given an $m \times n$ matrix $\widehat{{\mathbf M}}$, the prototypical RSVD algorithm outputs an approximation of the $k$ leading left singular vectors of $\widehat{\mathbf{M}}$ by computing the SVD of $\widehat{\mathbf{M}} (\widehat{{\mathbf M}}^{\top} \widehat{\mathbf{M}})^{g} \mathbf G$; here $g \geq 1$ is an integer and $\mathbf G \in \mathbb{R}^{n \times \widetilde{k}}$ is a random Gaussian sketching matrix with $\widetilde{k} \geq k$. In this paper we derive upper bounds for the $\ell_2$ and $\ell_{2,\infty}$ distances between the exact left singular vectors $\widehat{\mathbf{U}}$ of $\widehat{\mathbf{M}}$ and its approximation $\widehat{\mathbf{U}}_g$ (obtained via RSVD), as well as entrywise error bounds when $\widehat{\mathbf{M}}$ is projected onto $\widehat{\mathbf{U}}_g \widehat{\mathbf{U}}_g^{\top}$. These bounds depend on the singular values gap and number of power iterations $g$, and smaller gap requires larger values of $g$ to guarantee the convergences of the $\ell_2$ and $\ell_{2,\infty}$ distances. We apply our theoretical results to settings where $\widehat{\mathbf{M}}$ is an additive perturbation of some unobserved signal matrix $\mathbf{M}$. In particular, we obtain the nearly-optimal convergence rate and asymptotic normality for RSVD on three inference problems, namely, subspace estimation and community detection in random graphs, noisy matrix completion, and PCA with missing data.

math.ST↗

arXiv2021

Classification of high-dimensional data with spiked covariance matrix structure

We study the classification problem for high-dimensional data with $n$ observations on $p$ features where the $p \times p$ covariance matrix $\Sigma$ exhibits a spiked eigenvalue structure and the vector $\zeta$, given by the difference between the {\em whitened} mean vectors, is sparse. We analyze an adaptive classifier (adaptive with respect to the sparsity $s$) that first performs dimension reduction on the feature vectors prior to classification in the dimensionally reduced space, i.e., the classifier whitens the data, then screens the features by keeping only those corresponding to the $s$ largest coordinates of $\zeta$ and finally applies Fisher linear discriminant on the selected features. Leveraging recent results on entrywise matrix perturbation bounds for covariance matrices, we show that the resulting classifier is Bayes optimal whenever $n \rightarrow \infty$ and $s \sqrt{n^{-1} \ln p} \rightarrow 0$. Notably, our theory also guarantees Bayes optimality for the corresponding quadratic discriminant analysis (QDA). Experimental results on real and synthetic data further indicate that the proposed approach is competitive with state-of-the-art methods while operating on a substantially lower-dimensional representation.

stat.ML↗

arXiv2021

Popularity Adjusted Block Models are Generalized Random Dot Product Graphs

We connect two random graph models, the Popularity Adjusted Block Model (PABM) and the Generalized Random Dot Product Graph (GRDPG), by demonstrating that the PABM is a special case of the GRDPG in which communities correspond to mutually orthogonal subspaces of latent vectors. This insight allows us to construct new algorithms for community detection and parameter estimation for the PABM, as well as improve an existing algorithm that relies on Sparse Subspace Clustering. Using established asymptotic properties of Adjacency Spectral Embedding for the GRDPG, we derive asymptotic properties of these algorithms. In particular, we demonstrate that the absolute number of community detection errors tends to zero as the number of graph vertices tends to infinity. Simulation experiments illustrate these properties.

stat.ML↗

arXiv2021

Hypothesis Testing for Equality of Latent Positions in Random Graphs

We consider the hypothesis testing problem that two vertices $i$ and $j$ of a generalized random dot product graph have the same latent positions, possibly up to scaling. Special cases of this hypothesis test include testing whether two vertices in a stochastic block model or degree-corrected stochastic block model graph have the same block membership vectors, or testing whether two vertices in a popularity adjusted block model have the same community assignment. We propose several test statistics based on the empirical Mahalanobis distances between the $i$th and $j$th rows of either the adjacency or the normalized Laplacian spectral embedding of the graph. We show that, under mild conditions, these test statistics have limiting chi-square distributions under both the null and local alternative hypothesis, and we derived explicit expressions for the non-centrality parameters under the local alternative. Using these limit results, we address the model selection problems including choosing between the standard stochastic block model and its degree-corrected variant, and choosing between the ER model and stochastic block model. The effectiveness of our proposed tests are illustrated via both simulation studies and real data applications.

stat.ME↗

arXiv2021

Exact Recovery of Community Structures Using DeepWalk and Node2vec

Random-walk based network embedding algorithms like DeepWalk and node2vec are widely used to obtain Euclidean representation of the nodes in a network prior to performing downstream inference tasks. However, despite their impressive empirical performance, there is a lack of theoretical results explaining their large-sample behavior. In this paper, we study node2vec and DeepWalk through the perspective of matrix factorization. In particular, we analyze these algorithms in the setting of community detection for stochastic blockmodel graphs (and their degree-corrected variants). By exploiting the row-wise uniform perturbation bound for leading singular vectors, we derive high-probability error bounds between the matrix factorization-based node2vec/DeepWalk embeddings and their true counterparts, uniformly over all node embeddings. Based on strong concentration results, we further show the perfect membership recovery by node2vec/DeepWalk, followed by $K$-means/medians algorithms. Specifically, as the network becomes sparser, our results guarantee that with large enough window size and vertices number, applying $K$-means/medians on the matrix factorization-based node2vec embeddings can, with high probability, correctly recover the memberships of all vertices in a network generated from the stochastic blockmodel (or its degree-corrected variants). The theoretical justifications are mirrored in the numerical experiments and real data applications, for both the original node2vec and its matrix factorization variant.

stat.ML↗

arXiv2020

Nonparametric two-sample hypothesis testing for low-rank random graphs of differing sizes

Given two networks of differing sizes, it is of interest to test whether the two networks belong to the same distribution. We formalize the notion of "equality of distribution" under the framework of the generalized random dot product graph, which considers as special cases a number of popular network models with low-rank expectations. We then propose a nonparametric two-sample test statistic to conduct this test, assuming only that the networks have independent edges generated from low-rank probability matrices. Our proposed test statistic involves using the maximum mean discrepancy applied to suitably rotated rows of a graph embedding, where the rotation is estimated using optimal transport. We show that our test statistic, appropriately scaled, is consistent for sufficiently dense graphs, and we study its convergence under different sparsity regimes, and our results are demonstrated in numerical simulations.

math.ST↗

arXiv2020

Vertex nomination between graphs via spectral embedding and quadratic programming

Given a network and a subset of interesting vertices whose identities are only partially known, the vertex nomination problem seeks to rank the remaining vertices in such a way that the interesting vertices are ranked at the top of the list. An important variant of this problem is vertex nomination in the multi-graphs setting. Given two graphs $G_1, G_2$ with common vertices and a vertex of interest $x \in G_1$, we wish to rank the vertices of $G_2$ such that the vertices most similar to $x$ are ranked at the top of the list. The current paper addresses this problem and proposes a method that first applies adjacency spectral graph embedding to embed the graphs into a common Euclidean space, and then solves a penalized linear assignment problem to obtain the nomination lists. Since the spectral embedding of the graphs are only unique up to orthogonal transformations, we present two approaches to eliminate this potential non-identifiability. One approach is based on orthogonal Procrustes and is applicable when there are enough vertices with known correspondence between the two graphs. Another approach uses adaptive point set registration and is applicable when there are few or no vertices with known correspondence. We show that our nomination scheme leads to accurate nomination under a generative model for pairs of random graphs that are approximately low-rank and possibly with pairwise edge correlations. We illustrate our algorithm's performance through simulation studies on synthetic data as well as analysis of a high-school friendship network and analysis of transition rates between web pages on the Bing search engine.

cs.SI↗

arXiv2020

Two-sample Testing on Latent Distance Graphs With Unknown Link Functions

We propose a valid and consistent test for the hypothesis that two latent distance random graphs on the same vertex set have the same generating latent positions, up to some unidentifiable similarity transformations. Our test statistic is based on first estimating the edge probabilities matrices by truncating the singular value decompositions of the averaged adjacency matrices in each population and then computing a Spearman rank correlation coefficient between these estimates. Experimental results on simulated data indicate that the test procedure has power even when there is only one sample from each population, provided that the number of vertices is not too small. Application on a dataset of neural connectome graphs showed that we can distinguish between scans from different age groups while application on a dataset of epileptogenic recordings showed that we can discriminate between seizure and non-seizure events.

stat.ME↗

arXiv2020

Learning 1-Dimensional Submanifolds for Subsequent Inference on Random Dot Product Graphs

A random dot product graph (RDPG) is a generative model for networks in which vertices correspond to positions in a latent Euclidean space and edge probabilities are determined by the dot products of the latent positions. We consider RDPGs for which the latent positions are randomly sampled from an unknown $1$-dimensional submanifold of the latent space. In principle, restricted inference, i.e., procedures that exploit the structure of the submanifold, should be more effective than unrestricted inference; however, it is not clear how to conduct restricted inference when the submanifold is unknown. We submit that techniques for manifold learning can be used to learn the unknown submanifold well enough to realize benefit from restricted inference. To illustrate, we test $1$- and $2$-sample hypotheses about the Fr\'{e}chet means of small communities of vertices, using the complete set of vertices to infer latent structure. We propose test statistics that deploy the Isomap procedure for manifold learning, using shortest path distances on neighborhood graphs constructed from estimated latent positions to estimate arc lengths on the unknown $1$-dimensional submanifold. Unlike conventional applications of Isomap, the estimated latent positions do not lie on the submanifold of interest. We extend existing convergence results for Isomap to this setting and use them to demonstrate that, as the number of auxiliary vertices increases, the power of our test converges to the power of the corresponding test when the submanifold is known. Finally, we apply our methods to an inference problem that arises in studying the connectome of the Drosophila larval mushroom body. The univariate learnt manifold test rejects ($p<0.05$), while the multivariate ambient space test does not ($p\gg0.05$), illustrating the value of identifying and exploiting low-dimensional structure for subsequent inference.

stat.ML↗