In this paper we simulate an ensemble of cooperating, mobile sensing agents that implement the cyclic stochastic optimization (CSO) algorithm in an attempt to survey and track multiple targets. In the CSO algorithm proposed, each agent uses its sensed measurements, its shared information, and its predictions of others' future motion to decide on its next action. This decision is selected to minimize a loss function that decreases as the uncertainty in the targets' state estimates decreases. Only noisy measurements of this loss function are available to each agent, and in this study, each agent attempts to minimize this function by calculating its stochastic gradient. This paper examines, via simulation-based experiments, the implications and applicability of CSO convergence in three dimensions.

Finite Mixture Regression (FMR) refers to the mixture modeling scheme which learns multiple regression models from the training data set. Each of them is in charge of a subset. FMR is an effective scheme for handling sample heterogeneity, where a single regression model is not enough for capturing the complexities of the conditional distribution of the observed samples given the features. In this paper, we propose an FMR model that 1) finds sample clusters and jointly models multiple incomplete mixed-type targets simultaneously, 2) achieves shared feature selection among tasks and cluster components, and 3) detects anomaly tasks or clustered structure among tasks, and accommodates outlier samples. We provide non-asymptotic oracle performance bounds for our model under a high-dimensional learning framework. The proposed model is evaluated on both synthetic and real-world data sets. The results show that our model can achieve state-of-the-art performance.

Conditional estimation given specific covariate values (i.e., local conditional estimation or functional estimation) is ubiquitously useful with applications in engineering, social and natural sciences. Existing data-driven non-parametric estimators mostly focus on structured homogeneous data (e.g., weakly independent and stationary data), thus they are sensitive to adversarial noise and may perform poorly under a low sample size. To alleviate these issues, we propose a new distributionally robust estimator that generates non-parametric local estimates by minimizing the worst-case conditional expected loss over all adversarial distributions in a Wasserstein ambiguity set. We show that despite being generally intractable, the local estimator can be efficiently found via convex optimization under broadly applicable settings, and it is robust to the corruption and heterogeneity of the data. Experiments with synthetic and MNIST datasets show the competitive performance of this new class of estimators.

We consider the parameter estimation problem of a probabilistic generative model prescribed using a natural exponential family of distributions. For this problem, the typical maximum likelihood estimator usually overfits under limited training sample size, is sensitive to noise and may perform poorly on downstream predictive tasks. To mitigate these issues, we propose a distributionally robust maximum likelihood estimator that minimizes the worst-case expected log-loss uniformly over a parametric Kullback-Leibler ball around a parametric nominal distribution. Leveraging the analytical expression of the Kullback-Leibler divergence between two distributions in the same natural exponential family, we show that the min-max estimation problem is tractable in a broad setting, including the robust training of generalized linear models. Our novel robust estimator also enjoys statistical consistency and delivers promising empirical results in both regression and classification tasks.

In this paper we propose a new method to learn the underlying acyclic mixed graph of a linear non-Gaussian structural equation model given observational data. We build on an algorithm proposed by Wang and Drton, and we show that one can augment the hidden variable structure of the recovered model by learning {\em multidirected edges} rather than only directed and bidirected ones. Multidirected edges appear when more than two of the observed variables have a hidden common cause. We detect the presence of such hidden causes by looking at higher order cumulants and exploiting the multi-trek rule. Our method recovers the correct structure when the underlying graph is a bow-free acyclic mixed graph with potential multi-directed edges.

It is well understood that Dynamic Time Warping (DTW) is effective in revealing similarities between time series that do not align perfectly. In this paper, we illustrate this on spectroscopy time-series data. We show that DTW is effective in improving accuracy on a regression task when only a single wavelength is considered. When combined with k-Nearest Neighbour, DTW has the added advantage that it can reveal similarities and differences between samples at the level of the time-series. However, in the problem, we consider here data is available across a spectrum of wavelengths. If aggregate statistics (means, variances) are used across many wavelengths the benefits of DTW are no longer apparent. We present this as another example of a situation where big data trumps sophisticated models in Machine Learning.

Sampling is a fundamental and arguably very important task with numerous applications in Machine Learning. One approach to sample from a high dimensional distribution $e^{-f}$ for some function $f$ is the Langevin Algorithm (LA). Recently, there has been a lot of progress in showing fast convergence of LA even in cases where $f$ is non-convex, notably [53], [39] in which the former paper focuses on functions $f$ defined in $\mathbb{R}^n$ and the latter paper focuses on functions with symmetries (like matrix completion type objectives) with manifold structure. Our work generalizes the results of [53] where $f$ is defined on a manifold $M$ rather than $\mathbb{R}^n$. From technical point of view, we show that KL decreases in a geometric rate whenever the distribution $e^{-f}$ satisfies a log-Sobolev inequality on $M$.

Due to lack of data, overfitting ubiquitously exists in real-world applications of deep neural networks (DNNs). In this paper, we propose advanced dropout, a model-free methodology, to mitigate overfitting and improve the performance of DNNs. The advanced dropout technique applies a model-free and easily implemented distribution with a parametric prior, and adaptively adjusts dropout rate. Specifically, the distribution parameters are optimized by stochastic gradient variational Bayes (SGVB) inference in order to carry out an end-to-end training of DNNs. We evaluate the effectiveness of the advanced dropout against nine dropout techniques on five widely used datasets in computer vision. The advanced dropout outperforms all the referred techniques by 0.83% on average for all the datasets. An ablation study is conducted to analyze the effectiveness of each component. Meanwhile, convergence of dropout rate and ability to prevent overfitting are discussed in terms of classification performance. Moreover, we extend the application of the advanced dropout to uncertainty inference and network pruning, and we find that the advanced dropout is superior to the corresponding referred methods. The advanced dropout improves classification accuracies by 4% in uncertainty inference and by 0.2% and 0.5% when pruning more than 90% of nodes and 99.8% of parameters, respectively.

Modern machine learning methods are often overparametrized, allowing adaptation to the data at a fine level. This can seem puzzling; in the worst case, such models do not need to generalize. This puzzle inspired a great amount of work, arguing when overparametrization reduces test error, in a phenomenon called "double descent". Recent work aimed to understand in greater depth why overparametrization is helpful for generalization. This leads to discovering the unimodality of variance as a function of the level of parametrization, and to decomposing the variance into that arising from label noise, initialization, and randomness in the training data to understand the sources of the error. In this work we develop a deeper understanding of this area. Specifically, we propose using the analysis of variance (ANOVA) to decompose the variance in the test error in a symmetric way, for studying the generalization performance of certain two-layer linear and non-linear networks. The advantage of the analysis of variance is that it reveals the effects of initialization, label noise, and training data more clearly than prior approaches. Moreover, we also study the monotonicity and unimodality of the variance components. While prior work studied the unimodality of the overall variance, we study the properties of each term in variance decomposition. One key insight is that in typical settings, the interaction between training samples and initialization can dominate the variance; surprisingly being larger than their marginal effect. Also, we characterize "phase transitions" where the variance changes from unimodal to monotone. On a technical level, we leverage advanced deterministic equivalent techniques for Haar random matrices, that---to our knowledge---have not yet been used in the area. We also verify our results in numerical simulations and on empirical data examples.

This paper looks into the problem of handling imbalanced data in a multi-label classification problem. The problem is solved by proposing two novel methods that primarily exploit the geometric relationship between the feature vectors. The first one is an undersampling algorithm that uses angle between feature vectors to select more informative samples while rejecting the less informative ones. A suitable criterion is proposed to define the informativeness of a given sample. The second one is an oversampling algorithm that uses a generative algorithm to create new synthetic data that respects all class boundaries. This is achieved by finding \emph{no man's land} based on Euclidean distance between the feature vectors. The efficacy of the proposed methods is analyzed by solving a generic multi-class recognition problem based on mixture of Gaussians. The superiority of the proposed algorithms is established through comparison with other state-of-the-art methods, including SMOTE and ADASYN, over ten different publicly available datasets exhibiting high-to-extreme data imbalance. These two methods are combined into a single data processing framework and is labeled as ``GICaPS'' to highlight the role of geometry-based information (GI) sampling and Class-Prioritized Synthesis (CaPS) in dealing with multi-class data imbalance problem, thereby making a novel contribution in this field.