arXiv subjects

Jianqing Fan

Publications and source records attributed to Jianqing Fan.

At least 127 records · Page 7Linked to original sources

arXiv2019

Low-Rank Principal Eigenmatrix Analysis

Sparse PCA is a widely used technique for high-dimensional data analysis. In this paper, we propose a new method called low-rank principal eigenmatrix analysis. Different from sparse PCA, the dominant eigenvectors are allowed to be dense but are assumed to have a low-rank structure when matricized appropriately. Such a structure arises naturally in several practical cases: Indeed the top eigenvector of a circulant matrix, when matricized appropriately is a rank-1 matrix. We propose a matricized rank-truncated power method that could be efficiently implemented and establish its computational and statistical properties. Extensive experiments on several synthetic data sets demonstrate the competitive empirical performance of our method.

stat.ML↗

arXiv2019

A Selective Overview of Deep Learning

Deep learning has arguably achieved tremendous success in recent years. In simple words, deep learning uses the composition of many nonlinear functions to model the complex dependency between input features and labels. While neural networks have a long history, recent advances have greatly improved their performance in computer vision, natural language processing, etc. From the statistical and scientific perspective, it is natural to ask: What is deep learning? What are the new characteristics of deep learning, compared with classical methods? What are the theoretical foundations of deep learning? To answer these questions, we introduce common neural network models (e.g., convolutional neural nets, recurrent neural nets, generative adversarial nets) and training techniques (e.g., stochastic gradient descent, dropout, batch normalization) from a statistical point of view. Along the way, we highlight new characteristics of deep learning (including depth and over-parametrization) and explain their practical and theoretical benefits. We also sample recent results on theories of deep learning, many of which are only suggestive. While a complete understanding of deep learning remains elusive, we hope that our perspectives and discussions serve as a stimulus for new statistical research.

stat.ML↗

arXiv2019

Gradient Descent with Random Initialization: Fast Global Convergence for Nonconvex Phase Retrieval

This paper considers the problem of solving systems of quadratic equations, namely, recovering an object of interest $\mathbf{x}^{\natural}\in\mathbb{R}^{n}$ from $m$ quadratic equations/samples $y_{i}=(\mathbf{a}_{i}^{\top}\mathbf{x}^{\natural})^{2}$, $1\leq i\leq m$. This problem, also dubbed as phase retrieval, spans multiple domains including physical sciences and machine learning. We investigate the efficiency of gradient descent (or Wirtinger flow) designed for the nonconvex least squares problem. We prove that under Gaussian designs, gradient descent --- when randomly initialized --- yields an $ε$-accurate solution in $O\big(\log n+\log(1/ε)\big)$ iterations given nearly minimal samples, thus achieving near-optimal computational and sample complexities at once. This provides the first global convergence guarantee concerning vanilla gradient descent for phase retrieval, without the need of (i) carefully-designed initialization, (ii) sample splitting, or (iii) sophisticated saddle-point escaping schemes. All of these are achieved by exploiting the statistical models in analyzing optimization algorithms, via a leave-one-out approach that enables the decoupling of certain statistical dependency between the gradient descent iterates and the data.

stat.ML↗

arXiv2019

Estimation of False Discovery Proportion with Unknown Dependence

Large-scale multiple testing with highly correlated test statistics arises frequently in many scientific research. Incorporating correlation information in estimating false discovery proportion has attracted increasing attention in recent years. When the covariance matrix of test statistics is known, Fan, Han & Gu (2012) provided a consistent estimate of False Discovery Proportion (FDP) under arbitrary dependence structure. However, the covariance matrix is often unknown in many applications and such dependence information has to be estimated before estimating FDP (Efron, 2010). The estimation accuracy can greatly affect the convergence result of FDP or even violate its consistency. In the current paper, we provide methodological modification and theoretical investigations for estimation of FDP with unknown covariance. First we develop requirements for estimates of eigenvalues and eigenvectors such that we can obtain a consistent estimate of FDP. Secondly we give conditions on the dependence structures such that the estimate of FDP is consistent. Such dependence structures include sparse covariance matrices, which have been popularly considered in the contemporary random matrix theory. When data are sampled from an approximate factor model, which encompasses most practical situations, we provide a consistent estimate of FDP via exploiting this specific dependence structure. The results are further demonstrated by simulation studies and some real data applications.

stat.ME↗

arXiv2019

Bayesian Factor-adjusted Sparse Regression

This paper investigates the high-dimensional linear regression with highly correlated covariates. In this setup, the traditional sparsity assumption on the regression coefficients often fails to hold, and consequently many model selection procedures do not work. To address this challenge, we model the variations of covariates by a factor structure. Specifically, strong correlations among covariates are explained by common factors and the remaining variations are interpreted as idiosyncratic components of each covariate. This leads to a factor-adjusted regression model with both common factors and idiosyncratic components as covariates. We generalize the traditional sparsity assumption accordingly and assume that all common factors but only a small number of idiosyncratic components contribute to the response. A Bayesian procedure with a spike-and-slab prior is then proposed for parameter estimation and model selection. Simulation studies show that our Bayesian method outperforms its lasso analogue, manifests insensitivity to the overestimates of the number of common factors, pays a negligible price in the no correlation case, and scales up well with increasing sample size, dimensionality and sparsity. Numerical results on a real dataset of U.S. bond risk premia and macroeconomic indicators lend strong support to our methodology.

stat.ME↗

arXiv2019

A Projection Based Conditional Dependence Measure with Applications to High-dimensional Undirected Graphical Models

Measuring conditional dependence is an important topic in statistics with broad applications including graphical models. Under a factor model setting, a new conditional dependence measure based on projection is proposed. The corresponding conditional independence test is developed with the asymptotic null distribution unveiled where the number of factors could be high-dimensional. It is also shown that the new test has control over the asymptotic significance level and can be calculated efficiently. A generic method for building dependency graphs without Gaussian assumption using the new test is elaborated. Numerical results and real data analysis show the superiority of the new method.

stat.ME↗

arXiv2018

Higher Moment Estimation for Elliptically-distributed Data: Is it Necessary to Use a Sledgehammer to Crack an Egg?

Multivariate elliptically-contoured distributions are widely used for modeling economic and financial data. We study the problem of estimating moment parameters of a semi-parametric elliptical model in a high-dimensional setting. Such estimators are useful for financial data analysis and quadratic discriminant analysis. For low-dimensional elliptical models, efficient moment estimators can be obtained by plugging in an estimate of the precision matrix. Natural generalizations of the plug-in estimator to high-dimensional settings perform unsatisfactorily, due to estimating a large precision matrix. Do we really need a sledgehammer to crack an egg? Fortunately, we discover that moment parameters can be efficiently estimated without estimating the precision matrix in high-dimension. We propose a marginal aggregation estimator (MAE) for moment parameters. The MAE only requires estimating the diagonal of covariance matrix and is convenient to implement. With mild sparsity on the covariance structure, we prove that the asymptotic variance of MAE is the same as the ideal plug-in estimator which knows the true precision matrix, so MAE is asymptotically efficient. We also extend MAE to a block-wise aggregation estimator (BAE) when estimates of diagonal blocks of covariance matrix are available. The performance of our methods is validated by extensive simulations and an application to financial returns.

stat.ME↗

arXiv2018

Adaptive Huber Regression

Big data can easily be contaminated by outliers or contain variables with heavy-tailed distributions, which makes many conventional methods inadequate. To address this challenge, we propose the adaptive Huber regression for robust estimation and inference. The key observation is that the robustification parameter should adapt to the sample size, dimension and moments for optimal tradeoff between bias and robustness. Our theoretical framework deals with heavy-tailed distributions with bounded $(1+δ)$-th moment for any $δ> 0$. We establish a sharp phase transition for robust estimation of regression parameters in both low and high dimensions: when $δ\geq 1$, the estimator admits a sub-Gaussian-type deviation bound without sub-Gaussian assumptions on the data, while only a slower rate is available in the regime $0<δ< 1$. Furthermore, this transition is smooth and optimal. In addition, we extend the methodology to allow both heavy-tailed predictors and observation noise. Simulation studies lend further support to the theory. In a genetic study of cancer cell lines that exhibit heavy-tailedness, the proposed methods are shown to be more robust and predictive.

math.ST↗

arXiv2018

FarmTest: Factor-Adjusted Robust Multiple Testing with Approximate False Discovery Control

Large-scale multiple testing with correlated and heavy-tailed data arises in a wide range of research areas from genomics, medical imaging to finance. Conventional methods for estimating the false discovery proportion (FDP) often ignore the effect of heavy-tailedness and the dependence structure among test statistics, and thus may lead to inefficient or even inconsistent estimation. Also, the commonly imposed joint normality assumption is arguably too stringent for many applications. To address these challenges, in this paper we propose a Factor-Adjusted Robust Multiple Testing (FarmTest) procedure for large-scale simultaneous inference with control of the false discovery proportion. We demonstrate that robust factor adjustments are extremely important in both controlling the FDP and improving the power. We identify general conditions under which the proposed method produces consistent estimate of the FDP. As a byproduct that is of independent interest, we establish an exponential-type deviation inequality for a robust $U$-type covariance estimator under the spectral norm. Extensive numerical experiments demonstrate the advantage of the proposed method over several state-of-the-art methods especially when the data are generated from heavy-tailed distributions. The proposed procedures are implemented in the R-package FarmTest.

stat.ME↗

arXiv2018

Augmented Factor Models with Applications to Validating Market Risk Factors and Forecasting Bond Risk Premia

We study factor models augmented by observed covariates that have explanatory powers on the unknown factors. In financial factor models, the unknown factors can be reasonably well explained by a few observable proxies, such as the Fama-French factors. In diffusion index forecasts, identified factors are strongly related to several directly measurable economic variables such as consumption-wealth variable, financial ratios, and term spread. With those covariates, both the factors and loadings are identifiable up to a rotation matrix even only with a finite dimension. To incorporate the explanatory power of these covariates, we propose a smoothed principal component analysis (PCA): (i) regress the data onto the observed covariates, and (ii) take the principal components of the fitted data to estimate the loadings and factors. This allows us to accurately estimate the percentage of both explained and unexplained components in factors and thus to assess the explanatory power of covariates. We show that both the estimated factors and loadings can be estimated with improved rates of convergence compared to the benchmark method. The degree of improvement depends on the strength of the signals, representing the explanatory power of the covariates on the factors. The proposed estimator is robust to possibly heavy-tailed distributions. We apply the model to forecast US bond risk premia, and find that the observed macroeconomic characteristics contain strong explanatory powers of the factors. The gain of forecast is more substantial when the characteristics are incorporated to estimate the common factors than directly used for forecasts.

stat.ME↗

arXiv2018

Factor-Adjusted Regularized Model Selection

This paper studies model selection consistency for high dimensional sparse regression when data exhibits both cross-sectional and serial dependency. Most commonly-used model selection methods fail to consistently recover the true model when the covariates are highly correlated. Motivated by econometric studies, we consider the case where covariate dependence can be reduced through factor model, and propose a consistent strategy named Factor-Adjusted Regularized Model Selection (FarmSelect). By separating the latent factors from idiosyncratic components, we transform the problem from model selection with highly correlated covariates to that with weakly correlated variables. Model selection consistency as well as optimal rates of convergence are obtained under mild conditions. Numerical studies demonstrate the nice finite sample performance in terms of both model selection and out-of-sample prediction. Moreover, our method is flexible in a sense that it pays no price for weakly correlated and uncorrelated cases. Our method is applicable to a wide range of high dimensional sparse regression problems. An R-package FarmSelect is also provided for implementation.

stat.ME↗

arXiv2018

Curse of Heterogeneity: Computational Barriers in Sparse Mixture Models and Phase Retrieval

We study the fundamental tradeoffs between statistical accuracy and computational tractability in the analysis of high dimensional heterogeneous data. As examples, we study sparse Gaussian mixture model, mixture of sparse linear regressions, and sparse phase retrieval model. For these models, we exploit an oracle-based computational model to establish conjecture-free computationally feasible minimax lower bounds, which quantify the minimum signal strength required for the existence of any algorithm that is both computationally tractable and statistically accurate. Our analysis shows that there exist significant gaps between computationally feasible minimax risks and classical ones. These gaps quantify the statistical price we must pay to achieve computational tractability in the presence of data heterogeneity. Our results cover the problems of detection, estimation, support recovery, and clustering, and moreover, resolve several conjectures of Azizyan et al. (2013, 2015); Verzelen and Arias-Castro (2017); Cai et al. (2016). Interestingly, our results reveal a new but counter-intuitive phenomenon in heterogeneous data analysis that more data might lead to less computation complexity.

math.ST↗

arXiv2018

Robust high dimensional factor models with applications to statistical machine learning

Factor models are a class of powerful statistical models that have been widely used to deal with dependent measurements that arise frequently from various applications from genomics and neuroscience to economics and finance. As data are collected at an ever-growing scale, statistical machine learning faces some new challenges: high dimensionality, strong dependence among observed variables, heavy-tailed variables and heterogeneity. High-dimensional robust factor analysis serves as a powerful toolkit to conquer these challenges. This paper gives a selective overview on recent advance on high-dimensional factor models and their applications to statistics including Factor-Adjusted Robust Model selection (FarmSelect) and Factor-Adjusted Robust Multiple testing (FarmTest). We show that classical methods, especially principal component analysis (PCA), can be tailored to many new problems and provide powerful tools for statistical estimation and inference. We highlight PCA and its connections to matrix perturbation theory, robust statistics, random projection, false discovery rate, etc., and illustrate through several applications how insights from these fields yield solutions to modern challenges. We also present far-reaching connections between factor models and popular statistical learning problems, including network analysis and low-rank matrix recovery.

stat.ME↗

arXiv2018

Hoeffding's lemma for Markov Chains and its applications to statistical learning

We extend Hoeffding's lemma to general-state-space and not necessarily reversible Markov chains. Let $\{X_i\}_{i \ge 1}$ be a stationary Markov chain with invariant measure $π$ and absolute spectral gap $1-λ$, where $λ$ is defined as the operator norm of the transition kernel acting on mean zero and square-integrable functions with respect to $π$. Then, for any bounded functions $f_i: x \mapsto [a_i,b_i]$, the sum of $f_i(X_i)$ is sub-Gaussian with variance proxy $\frac{1+λ}{1-λ} \cdot \sum_i \frac{(b_i-a_i)^2}{4}$. This result differs from the classical Hoeffding's lemma by a multiplicative coefficient of $(1+λ)/(1-λ)$, and simplifies to the latter when $λ= 0$. The counterpart of Hoeffding's inequality for Markov chains immediately follows. Our results assume none of countable state space, reversibility and time-homogeneity of Markov chains and cover time-dependent functions with various ranges. We illustrate the utility of these results by applying them to six problems in statistics and machine learning.

math.ST↗

arXiv2018

Tensor Methods for Additive Index Models under Discordance and Heterogeneity

Motivated by the sampling problems and heterogeneity issues common in high- dimensional big datasets, we consider a class of discordant additive index models. We propose method of moments based procedures for estimating the indices of such discordant additive index models in both low and high-dimensional settings. Our estimators are based on factorizing certain moment tensors and are also applicable in the overcomplete setting, where the number of indices is more than the dimensionality of the datasets. Furthermore, we provide rates of convergence of our estimator in both high and low-dimensional setting. Establishing such results requires deriving tensor operator norm concentration inequalities that might be of independent interest. Finally, we provide simulation results supporting our theory. Our contributions extend the applicability of tensor methods for novel models in addition to making progress on understanding theoretical properties of such tensor methods.

math.ST↗

arXiv2018

Spectral Method and Regularized MLE Are Both Optimal for Top-$K$ Ranking

This paper is concerned with the problem of top-$K$ ranking from pairwise comparisons. Given a collection of $n$ items and a few pairwise comparisons across them, one wishes to identify the set of $K$ items that receive the highest ranks. To tackle this problem, we adopt the logistic parametric model --- the Bradley-Terry-Luce model, where each item is assigned a latent preference score, and where the outcome of each pairwise comparison depends solely on the relative scores of the two items involved. Recent works have made significant progress towards characterizing the performance (e.g. the mean square error for estimating the scores) of several classical methods, including the spectral method and the maximum likelihood estimator (MLE). However, where they stand regarding top-$K$ ranking remains unsettled. We demonstrate that under a natural random sampling model, the spectral method alone, or the regularized MLE alone, is minimax optimal in terms of the sample complexity --- the number of paired comparisons needed to ensure exact top-$K$ identification, for the fixed dynamic range regime. This is accomplished via optimal control of the entrywise error of the score estimates. We complement our theoretical studies by numerical experiments, confirming that both methods yield low entrywise errors for estimating the underlying scores. Our theory is established via a novel leave-one-out trick, which proves effective for analyzing both iterative and non-iterative procedures. Along the way, we derive an elementary eigenvector perturbation bound for probability transition matrices, which parallels the Davis-Kahan $\sinΘ$ theorem for symmetric matrices. This also allows us to close the gap between the $\ell_2$ error upper bound for the spectral method and the minimax lower limit.

stat.ML↗

arXiv2018

Optimal Subspace Estimation Using Overidentifying Vectors via Generalized Method of Moments

Many statistical models seek relationship between variables via subspaces of reduced dimensions. For instance, in factor models, variables are roughly distributed around a low dimensional subspace determined by the loading matrix; in mixed linear regression models, the coefficient vectors for different mixtures form a subspace that captures all regression functions; in multiple index models, the effect of covariates is summarized by the effective dimension reduction space. Such subspaces are typically unknown, and good estimates are crucial for data visualization, dimension reduction, diagnostics and estimation of unknown parameters. Usually, we can estimate these subspaces by computing moments from data. Often, there are many ways to estimate a subspace, by using moments of different orders, transformed moments, etc. A natural question is: how can we combine all these moment conditions and achieve optimality for subspace estimation? In this paper, we formulate our problem as estimation of an unknown subspace $\mathcal{S}$ of dimension $r$, given a set of overidentifying vectors $\{ \mathrm{\bf v}_\ell \}_{\ell=1}^m$ (namely $m \ge r$) that satisfy $\mathbb{E} \mathrm{\bf v}_{\ell} \in \mathcal{S}$ and have the form $$ \mathrm{\bf v}_\ell = \frac{1}{n} \sum_{i=1}^n \mathrm{\bf f}_\ell(\mathbf{x}_i, y_i), $$ where data are i.i.d. and each function $\mathrm{\bf f}_\ell$ is known. By exploiting certain covariance information related to $\mathrm{\bf v}_\ell$, our estimator of $\mathcal{S}$ uses an optimal weighting matrix and achieves the smallest asymptotic error, in terms of canonical angles. The analysis is based on the generalized method of moments that is tailored to our problem. Our method is applied to aforementioned models and distributed estimation of heterogeneous datasets, and may be potentially extended to analyze matrix completion, neural nets, among others.

stat.ME↗

arXiv2018

Distributed Estimation of Principal Eigenspaces

Principal component analysis (PCA) is fundamental to statistical machine learning. It extracts latent principal factors that contribute to the most variation of the data. When data are stored across multiple machines, however, communication cost can prohibit the computation of PCA in a central location and distributed algorithms for PCA are thus needed. This paper proposes and studies a distributed PCA algorithm: each node machine computes the top $K$ eigenvectors and transmits them to the central server; the central server then aggregates the information from all the node machines and conducts a PCA based on the aggregated information. We investigate the bias and variance for the resulting distributed estimator of the top $K$ eigenvectors. In particular, we show that for distributions with symmetric innovation, the empirical top eigenspaces are unbiased and hence the distributed PCA is "unbiased". We derive the rate of convergence for distributed PCA estimators, which depends explicitly on the effective rank of covariance, eigen-gap, and the number of machines. We show that when the number of machines is not unreasonably large, the distributed PCA performs as well as the whole sample PCA, even without full access of whole data. The theoretical results are verified by an extensive simulation study. We also extend our analysis to the heterogeneous case where the population covariance matrices are different across local machines but share similar top eigen-structures.

stat.CO↗