Sparse Group Lasso is a method of linear regression analysis that finds sparse parameters in terms of both feature groups and individual features. Block Coordinate Descent is a standard approach to obtain the parameters of Sparse Group Lasso, and iteratively updates the parameters for each parameter group. However, as an update of only one parameter group depends on all the parameter groups or data points, the computation cost is high when the number of the parameters or data points is large. This paper proposes a fast Block Coordinate Descent for Sparse Group Lasso. It efficiently skips the updates of the groups whose parameters must be zeros by using the parameters in one group. In addition, it preferentially updates parameters in a candidate group set, which contains groups whose parameters must not be zeros.

Linear regression is an important tool across many fields that work with sensitive human-sourced data. Significant prior work has focused on producing differentially private point estimates, which provide a privacy guarantee to individuals while still allowing modelers to draw insights from data by estimating regression coefficients. We investigate the problem of Bayesian linear regression, with the goal of computing posterior distributions that correctly quantify uncertainty given privately released statistics. We show that a naive approach that ignores the noise injected by the privacy mechanism does a poor job in realistic data settings. We then develop noise-aware methods that perform inference over the privacy mechanism and produce correct posteriors across a wide range of scenarios.

We consider training machine learning models using Training data located on multiple private and geographically-scattered servers with different privacy settings. Due to the distributed nature of the data, communicating with all collaborating private data owners simultaneously may prove challenging or altogether impossible. In this paper, we develop differentially-private asynchronous algorithms for collaboratively training machine-learning models on multiple private datasets. The asynchronous nature of the algorithms implies that a central learner interacts with the private data owners one-on-one whenever they are available for communication without needing to aggregate query responses to construct gradients of the entire fitness function. Therefore, the algorithm efficiently scales to many data owners. We define the cost of privacy as the difference between the fitness of a privacy-preserving machine-learning model and the fitness of trained machine-learning model in the absence of privacy concerns. We prove that we can forecast the performance of the proposed privacy-preserving asynchronous algorithms. We demonstrate that the cost of privacy has an upper bound that is inversely proportional to the combined size of the training datasets squared and the sum of the privacy budgets squared. We validate the theoretical results with experiments on financial and medical datasets. The experiments illustrate that collaboration among more than 10 data owners with at least 10,000 records with privacy budgets greater than or equal to 1 results in a superior machine-learning model in comparison to a model trained in isolation on only one of the datasets, illustrating the value of collaboration and the cost of the privacy. The number of the collaborating datasets can be lowered if the privacy budget is higher.

Spin glass models, such as the Sherrington-Kirkpatrick, Hopfield and Ising models, are all well-studied members of the exponential family of discrete distributions, and have been influential in a number of application domains where they are used to model correlation phenomena on networks. Conventionally these models have quadratic sufficient statistics and consequently capture correlations arising from pairwise interactions. In this work we study extensions of these to models with higher-order sufficient statistics, modeling behavior on a social network with peer-group effects. In particular, we model binary outcomes on a network as a higher-order spin glass, where the behavior of an individual depends on a linear function of their own vector of covariates and some polynomial function of the behavior of others, capturing peer-group effects. Using a {\em single}, high-dimensional sample from such model our goal is to recover the coefficients of the linear function as well as the strength of the peer-group effects. The heart of our result is a novel approach for showing strong concavity of the log pseudo-likelihood of the model, implying statistical error rate of $\sqrt{d/n}$ for the Maximum Pseudo-Likelihood Estimator (MPLE), where $d$ is the dimensionality of the covariate vectors and $n$ is the size of the network (number of nodes). Our model generalizes vanilla logistic regression as well as the peer-effect models studied in recent works, and our results extend these results to accommodate higher-order interactions.

As in other fields of artificial intelligence and prior to the emergence of Deep Learning, especially deep neural networks, artificial vision research was focused on a traditional Machine Learning approach. The traditional machine learning approach relies on developers massaging the data to extract the most salient or significant aspects from the data they are dealing with; that is, time sequences of frames, or videos. In this case, both scientific research and application development have been centered around identifying the most significant image elements that would allow, for example, facial and body recognition of the people who appear in the images, tracking them from one frame to another, or classifying the vehicles that move through a given area. After extracting this meaningful data, statistical methods are then employed to transform the representation into a so-called "understanding" of the real visual environment by using clustering, support-vector machines (SVMs), and filtering algorithms (linear, non-linear, regression), among others. This means that the merits of any given application lie in how well researchers and developers are able to source and generate data from the raw processed frames and transform it into useful structured data.

The sparse factorization of a large matrix is fundamental in modern statistical learning. In particular, the sparse singular value decomposition and its variants have been utilized in multivariate regression, factor analysis, biclustering, vector time series modeling, among others. The appeal of this factorization is owing to its power in discovering a highly-interpretable latent association network, either between samples and variables or between responses and predictors. However, many existing methods are either ad hoc without a general performance guarantee, or are computationally intensive, rendering them unsuitable for large-scale studies. We formulate the statistical problem as a sparse factor regression and tackle it with a divide-and-conquer approach. In the first stage of division, we consider both sequential and parallel approaches for simplifying the task into a set of co-sparse unit-rank estimation (CURE) problems, and establish the statistical underpinnings of these commonly-adopted and yet poorly understood deflation methods. In the second stage of division, we innovate a contended stagewise learning technique, consisting of a sequence of simple incremental updates, to efficiently trace out the whole solution paths of CURE. Our algorithm has a much lower computational complexity than alternating convex search, and the choice of the step size enables a flexible and principled tradeoff between statistical accuracy and computational efficiency. Our work is among the first to enable stagewise learning for non-convex problems, and the idea can be applicable in many multi-convex problems. Extensive simulation studies and an application in genetics demonstrate the effectiveness and scalability of our approach.

Machine learning has increasingly become the most used approach for inference and decision making in wearable sensor systems. However, recent studies have found that machine learning systems are easily fooled by the addition of adversarial perturbation to their inputs. What is more interesting is that the adversarial examples generated for one machine learning system can also degrade the performance of another. This property of adversarial examples is called transferability. In this work, we take the first strides in studying adversarial transferability in wearable sensor systems, from the following perspectives: 1) Transferability between machine learning models, 2) Transferability across subjects, 3) Transferability across sensor locations, and 4) Transferability across datasets. With Human Activity Recognition (HAR) as an example sensor system, we found strong untargeted transferability in all cases of transferability. Specifically, gradient-based attacks were able to achieve higher misclassification rates compared to non-gradient attacks. The misclassification rate of untargeted adversarial examples ranged from 20% to 98%. For targeted transferability between machine learning models, the success rate of adversarial examples was 100% for iterative attack methods. However, the success rate for other types of targeted transferability ranged from 20% to 0%. Our findings strongly suggest that adversarial transferability has serious consequences not only in sensor systems but also across the broad spectrum of ubiquitous computing.

Causal Inference methods based on observational data are an alternative for applications where collecting the counterfactual data or realizing a more standard experiment is not possible. In this work, our goal is to combine several observational causal inference methods to learn new causes in applications where some causes are well known. We validate the proposed method on The Cancer Genome Atlas (TCGA) dataset to identify genes that potentially cause metastasis.

When applying machine learning/statistical methods to the environmental sciences, nonlinear regression (NLR) models often perform only slightly better and occasionally worse than linear regression (LR). The proposed reason for this conundrum is that NLR models can give predictions much worse than LR when given input data which lie outside the domain used in model training. Continuous unbounded variables are widely used in environmental sciences, whence not uncommon for new input data to lie far outside the training domain. For six environmental datasets, inputs in the test data were classified as "outliers" and "non-outliers" based on the Mahalanobis distance from the training input data. The prediction scores (mean absolute error, Spearman correlation) showed NLR to outperform LR for the non-outliers, but often underperform LR for the outliers. An approach based on Occam's Razor (OR) was proposed, where linear extrapolation was used instead of nonlinear extrapolation for the outliers. The linear extrapolation to the outlier domain was based on the NLR model within the non-outlier domain. This NLR$_{\mathrm{OR}}$ approach reduced occurrences of very poor extrapolation by NLR, and it tended to outperform NLR and LR for the outliers. In conclusion, input test data should be screened for outliers. For outliers, the unreliable NLR predictions can be replaced by NLR$_{\mathrm{OR}}$ or LR predictions, or by issuing a "no reliable prediction" warning.

KernelExplainer:- Kernel SHAP is a method that uses a special weighted linear regression to compute the importance of each feature.