Support vector regression (SVR) has been widely used to reduce the high computational cost of computer simulation. SVR assumes the input parameters have equal sample sizes, but unequal sample sizes are often encountered in engineering practices. To solve this issue, a new prediction approach based on SVR, namely as high-low-level SVR approach (HL-SVR) is proposed for data modeling of input parameters of unequal sample sizes in this paper. The proposed approach is consisted of low-level SVR models for the input parameters of larger sample sizes and high-level SVR model for the input parameters of smaller sample sizes. For each training point of the input parameters of smaller sample sizes, one low-level SVR model is built based on its corresponding input parameters of larger sample sizes and their responses of interest. The high-level SVR model is built based on the obtained responses from the low-level SVR models and the input parameters of smaller sample sizes. Several numerical examples are used to validate the performance of HL-SVR. The experimental results indicate that HL-SVR can produce more accurate prediction results than conventional SVR. The proposed approach is applied on the stress analysis of dental implant, which the structural parameters have massive samples but the material of implant can only be selected from several Ti and its alloys. The prediction performance of the proposed approach is much better than the conventional SVR. The proposed approach can be used for the design, optimization and analysis of engineering systems with input parameters of unequal sample sizes.

With the rapid growth of data, distributed stochastic gradient descent~(DSGD) has been widely used for solving large-scale machine learning problems. Due to the latency and limited bandwidth of network, communication has become the bottleneck of DSGD when we need to train large scale models, like deep neural networks. Communication compression with sparsified gradient, abbreviated as \emph{sparse communication}, has been widely used for reducing communication cost in DSGD. Recently, there has appeared one method, called deep gradient compression~(DGC), to combine memory gradient and momentum SGD for sparse communication. DGC has achieved promising performance in practise. However, the theory about the convergence of DGC is lack. In this paper, we propose a novel method, called \emph{\underline{g}}lobal \emph{\underline{m}}omentum \emph{\underline{c}}ompression~(GMC), for sparse communication in DSGD. GMC also combines memory gradient and momentum SGD. But different from DGC which adopts local momentum, GMC adopts global momentum. We theoretically prove the convergence rate of GMC for both convex and non-convex problems. To the best of our knowledge, this is the first work that proves the convergence of distributed momentum SGD~(DMSGD) with sparse communication and memory gradient. Empirical results show that, compared with the DMSGD counterpart without sparse communication, GMC can reduce the communication cost by approximately 100 fold without loss of generalization accuracy. GMC can also achieve comparable~(sometimes better) performance compared with DGC, with extra theoretical guarantee.

Attributed graphs, which contain rich contextual features beyond just network structure, are ubiquitous and have been observed to benefit various network analytics applications. Graph structure optimization, aiming to find the optimal graphs in terms of some specific measures, has become an effective computational tool in complex network analysis. However, traditional model-free methods suffer from the expensive computational cost of evaluating graphs; existing vectorial Bayesian optimization methods cannot be directly applied to attributed graphs and have the scalability issue due to the use of Gaussian processes (GPs). To bridge the gap, in this paper, we propose a novel scalable Deep Graph Bayesian Optimization (DGBO) method on attributed graphs. The proposed DGBO prevents the cubical complexity of the GPs by adopting a deep graph neural network to surrogate black-box functions, and can scale linearly with the number of observations. Intensive experiments are conducted on both artificial and real-world problems, including molecular discovery and urban road network design, and demonstrate the effectiveness of the DGBO compared with the state-of-the-art.

We show that the classification performance of Graph Convolutional Networks is related to the alignment between features, graph and ground truth, which we quantify using a subspace alignment measure corresponding to the Frobenius norm of the matrix of pairwise chordal distances between three subspaces associated with features, graph and ground truth. The proposed measure is based on the principal angles between subspaces and has both spectral and geometrical interpretations. We showcase the relationship between the subspace alignment measure and the classification performance through the study of limiting cases of Graph Convolutional Networks as well as systematic randomizations of both features and graph structure applied to a constructive example and several examples of citation networks of different origin. The analysis also reveals the relative importance of the graph and features for classification purposes.

Distributed stochastic gradient descent~(DSGD) has been widely used for optimizing large-scale machine learning models, including both convex and non-convex models. With the rapid growth of model size, huge communication cost has been the bottleneck of traditional DSGD. Recently, many communication compression methods have been proposed. Memory-based distributed stochastic gradient descent~(M-DSGD) is one of the efficient methods since each worker communicates a sparse vector in each iteration so that the communication cost is small. Recent works propose the convergence rate of M-DSGD when it adopts vanilla SGD. However, there is still a lack of convergence theory for M-DSGD when it adopts momentum SGD. In this paper, we propose a universal convergence analysis for M-DSGD by introducing \emph{transformation equation}. The transformation equation describes the relation between traditional DSGD and M-DSGD so that we can transform M-DSGD to its corresponding DSGD. Hence we get the convergence rate of M-DSGD with momentum for both convex and non-convex problems. Furthermore, we combine M-DSGD and stagewise learning that the learning rate of M-DSGD in each stage is a constant and is decreased by stage, instead of iteration. Using the transformation equation, we propose the convergence rate of stagewise M-DSGD which bridges the gap between theory and practice.

We consider the problem of sampling from a target distribution which is \emph{not necessarily logconcave}. Non-asymptotic analysis results are established in a suitable Wasserstein-type distance of the Stochastic Gradient Langevin Dynamics (SGLD) algorithm, when the gradient is driven by even \emph{dependent} data streams. Our estimates are sharper and \emph{uniform} in the number of iterations, in contrast to those in previous studies.

We provide a differentially private algorithm for hypothesis selection. Given samples from an unknown probability distribution $P$ and a set of $m$ probability distributions $\mathcal{H}$, the goal is to output, in a $\varepsilon$-differentially private manner, a distribution from $\mathcal{H}$ whose total variation distance to $P$ is comparable to that of the best such distribution (which we denote by $\alpha$). The sample complexity of our basic algorithm is $O\left(\frac{\log m}{\alpha^2} + \frac{\log m}{\alpha \varepsilon}\right)$, representing a minimal cost for privacy when compared to the non-private algorithm. We also can handle infinite hypothesis classes $\mathcal{H}$ by relaxing to $(\varepsilon,\delta)$-differential privacy. We apply our hypothesis selection algorithm to give learning algorithms for a number of natural distribution classes, including Gaussians, product distributions, sums of independent random variables, piecewise polynomials, and mixture classes. Our hypothesis selection procedure allows us to generically convert a cover for a class to a learning algorithm, complementing known learning lower bounds which are in terms of the size of the packing number of the class. As the covering and packing numbers are often closely related, for constant $\alpha$, our algorithms achieve the optimal sample complexity for many classes of interest. Finally, we describe an application to private distribution-free PAC learning.

Machine learning algorithms typically rely on optimization subroutines and are well-known to provide very effective outcomes for many types of problems. Here, we flip the reliance and ask the reverse question: can machine learning algorithms lead to more effective outcomes for optimization problems? Our goal is to train machine learning methods to automatically improve the performance of optimization and signal processing algorithms. As a proof of concept, we use our approach to improve two popular data processing subroutines in data science: stochastic gradient descent and greedy methods in compressed sensing. We provide experimental results that demonstrate the answer is ``yes'', machine learning algorithms do lead to more effective outcomes for optimization problems, and show the future potential for this research direction.

We present a new PAC-Bayesian generalization bound. Standard bounds contain a $\sqrt{L_n \cdot \operatorname{KL}/n}$ complexity term which dominates unless $L_n$, the empirical error of the learning algorithm's randomized predictions, vanishes. We manage to replace $L_n$ by a term which vanishes in many more situations, essentially whenever the employed learning algorithm is sufficiently stable on the dataset at hand. Our new bound consistently beats state-of-the-art bounds both on a toy example and on UCI datasets (with large enough $n$). Theoretically, unlike existing bounds, our new bound can be expected to converge to $0$ faster whenever a Bernstein/Tsybakov condition holds, thus connecting PAC-Bayesian generalization and excess risk bounds --- for the latter it has long been known that faster convergence can be obtained under Bernstein conditions. Our main technical tool is a new concentration inequality which is like Bernstein's but with $X^2$ taken outside its expectation.

We propose a neural architecture search (NAS) algorithm, Petridish, to iteratively add shortcut connections to existing network layers. The added shortcut connections effectively perform gradient boosting on the augmented layers. The proposed algorithm is motivated by the feature selection algorithm forward stage-wise linear regression, since we consider NAS as a generalization of feature selection for regression, where NAS selects shortcuts among layers instead of selecting features. In order to reduce the number of trials of possible connection combinations, we train jointly all possible connections at each stage of growth while leveraging feature selection techniques to choose a subset of them. We experimentally show this process to be an efficient forward architecture search algorithm that can find competitive models using few GPU days in both the search space of repeatable network modules (cell-search) and the space of general networks (macro-search). Petridish is particularly well-suited for warm-starting from existing models crucial for lifelong-learning scenarios.