The maximum mean discrepancy and Wasserstein distance are popular distance measures between distributions and play important roles in many machine learning problems such as metric learning, generative modeling, domain adaption, and clustering. However, since they are functions of pair-wise distances between data points in two distributions, they do not exploit the potential manifold properties of data such as smoothness and hence are not effective in measuring the dissimilarity between the two distributions in the form of manifolds. In this paper, different from existing measures, we propose a novel distance called Mutual Regression Distance (MRD) induced by a constrained mutual regression problem, which can exploit the manifold property of data. We prove that MRD is a pseudometric that satisfies almost all the axioms of a metric. Since the optimization of the original MRD is costly, we provide a tight MRD and a simplified MRD, based on which a heuristic algorithm is established. We also provide kernel variants of MRDs that are more effective in handling nonlinear data. Our MRDs especially the simplified MRDs have much lower computational complexity than the Wasserstein distance. We provide theoretical guarantees, such as robustness, for MRDs. Finally, we apply MRDs to distribution clustering, generative models, and domain adaptation. The numerical results demonstrate the effectiveness and superiority of MRDs compared to the baselines.

Conditional independence (CI) testing is a fundamental task in statistics and machine learning, but its effectiveness is hindered by the challenges posed by high-dimensional conditioning variables and limited data samples. This article introduces a novel testing approach to address these challenges and enhance control of the type I error while achieving high power under alternative hypotheses. The proposed approach incorporates a computationally efficient classifier-based conditional mutual information (CMI) estimator, capable of capturing intricate dependence structures among variables. To approximate a distribution encoding the null hypothesis, a k -nearest-neighbor local sampling strategy is employed. An important advantage of this approach is its ability to operate without assumptions about distribution forms or feature dependencies. Furthermore, it eliminates the need to derive asymptotic null distributions for the estimated CMI and avoids dataset splitting, making it particularly suitable for small datasets.

Stochastic Gradient Descent (SGD) is arguably the most important single algorithm in modern machine learning. Although SGD with unbiased gradient estimators has been studied extensively over at least half a century, SGD variants relying on biased estimators are rare. Nevertheless, there has been an increased interest in this topic in recent years. However, existing literature on SGD with biased estimators lacks coherence since each new paper relies on a different set of assumptions, without any clear understanding of how they are connected, which may lead to confusion. We address this gap by establishing connections among the existing assumptions, and presenting a comprehensive map of the underlying relationships. Additionally, we introduce a new set of assumptions that is provably weaker than all previous assumptions, and use it to present a thorough analysis of BiasedSGD in both convex and non-convex settings, offering advantages over previous results.

We study the problem of uncertainty quantification for time series prediction, with the goal of providing easy-to-use algorithms with formal guarantees. The algorithms we present build upon ideas from conformal prediction and control theory, are able to prospectively model conformal scores in an online setting, and adapt to the presence of systematic errors due to seasonality, trends, and general distribution shifts. Our theory both simplifies and strengthens existing analyses in online conformal prediction. Experiments on 4-week-ahead forecasting of statewide COVID-19 death counts in the U.S. show an improvement in coverage over the ensemble forecaster used inofficial CDC communications. We also run experiments on predicting electricity demand, market returns, and temperature using autoregressive, Theta, Prophet, and Transformer models.

In past few years, several techniques have been proposed for training of linear Support Vector Machine (SVM) in limited-memory setting, where a dual block-coordinate descent (dual-BCD) method was used to balance cost spent on I/O and computation. In this paper, we consider the more general setting of regularized \emph{Empirical Risk Minimization (ERM)} when data cannot fit into memory. In particular, we generalize the existing block minimization framework based on strong duality and \emph{Augmented Lagrangian} technique to achieve global convergence for ERM with arbitrary convex loss function and regularizer. The block minimization framework is flexible in the sense that, given a solver working under sufficient memory, one can integrate it with the framework to obtain a solver globally convergent under limited-memory condition. We conduct experiments on L1-regularized classification and regression problems to corroborate our convergence theory and compare the proposed framework to algorithms adopted from online and distributed settings, which shows superiority of the proposed approach on data of size ten times larger than the memory capacity.

We study the canonical problem of linear regression under (\varepsilon,\delta) -differential privacy when the datapoints are sampled i.i.d. We provide the first provably efficient -- both computationally and statistically -- method for this problem, assuming standard assumptions on the data distribution. Our algorithm is a variant of the popular differentially private stochastic gradient descent (DP-SGD) algorithm with two key innovations: a full-batch gradient descent to improve sample complexity and a novel adaptive clipping to guarantee robustness. Our method requires only linear time in input size, and still matches the information theoretical optimal sample complexity up to a data distribution dependent condition number factor. Interestingly, the same algorithm, when applied to a setting where there is no adversarial corruption, still improves upon the existing state-of-the-art and achieves a near optimal sample complexity.

Consider the binary classification problem of predicting a target variable Y from a discrete feature vector X (X1,...,Xd). When the probability distribution P(X,Y) is known, the optimal classifier, leading to the minimum misclassification rate, is given by the Maximum A-posteriori Probability (MAP) decision rule. However, in practice, estimating the complete joint distribution P(X,Y) is computationally and statistically impossible for large values of d. Therefore, an alternative approach is to first estimate some low order marginals of the joint probability distribution P(X,Y) and then design the classifier based on the estimated low order marginals. This approach is also helpful when the complete training data instances are not available due to privacy concerns.

Over-parametrization is an important technique in training neural networks. In both theory and practice, training a larger network allows the optimization algorithm to avoid bad local optimal solutions. In this paper we study a closely related tensor decomposition problem: given an l -th order tensor in (R d) {\otimes l} of rank r (where r\ll d), can variants of gradient descent find a rank m decomposition where m r? We show that in a lazy training regime (similar to the NTK regime for neural networks) one needs at least m \Omega(d {l-1}), while a variant of gradient descent can find an approximate tensor when m O *(r {2.5l}\log d) . Our results show that gradient descent on over-parametrized objective could go beyond the lazy training regime and utilize certain low-rank structure in the data.

The weights and the parameter k \in \mathbb{N} regulate its bias-variance trade-off, and the trade-off implicitly affects the convergence rate of the excess risk for the k -NN classifier; several existing studies considered selecting optimal k and weights to obtain faster convergence rate. Whereas k -NN with non-negative weights has been developed widely, it was also proved that negative weights are essential for eradicating the bias terms and attaining optimal convergence rate. In this paper, we propose a novel multiscale k -NN (MS- k -NN), that extrapolates unweighted k -NN estimators from several k \ge 1 values to k 0, thus giving an imaginary 0-NN estimator. Our method implicitly computes optimal real-valued weights that are adaptive to the query and its neighbour points. We theoretically prove that the MS- k -NN attains the improved rate, which coincides with the existing optimal rate under some conditions.

Recent years have witnessed the superiority of non-convex sparse learning formulations over their convex counterparts in both theory and practice. However, due to the non-convexity and non-smoothness of the regularizer, how to efficiently solve the non-convex optimization problem for large-scale data is still quite challenging. Specifically, we develop a hybrid scheme which effectively integrates a Quasi-Newton (QN) step and a Gradient Descent (GD) step. Our contributions are as follows: (1) HONOR incorporates the second-order information to greatly speed up the convergence, while it avoids solving a regularized quadratic programming and only involves matrix-vector multiplications without explicitly forming the inverse Hessian matrix.