Many machine learning algorithms perform better when numerical input variables are scaled to a standard range. This includes algorithms that use a weighted sum of the input, like linear regression, and algorithms that use distance measures, like k-nearest neighbors. Standardizing is a popular scaling technique that subtracts the mean from values and divides by the standard deviation, transforming the probability distribution for an input variable to a standard Gaussian (zero mean and unit variance). Standardization can become skewed or biased if the input variable contains outlier values. To overcome this, the median and interquartile range can be used when standardizing numerical input variables, generally referred to as robust scaling.

In recent years, deep learning (DL) has contributed significantly to the improvement of motor-imagery brain-machine interfaces (MI-BMIs) based on electroencephalography(EEG). While achieving high classification accuracy, DL models have also grown in size, requiring a vast amount of memory and computational resources. This poses a major challenge to an embedded BMI solution that guarantees user privacy, reduced latency, and low power consumption by processing the data locally. In this paper, we propose EEG-TCNet, a novel temporal convolutional network (TCN) that achieves outstanding accuracy while requiring few trainable parameters. Its low memory footprint and low computational complexity for inference make it suitable for embedded classification on resource-limited devices at the edge. Experimental results on the BCI Competition IV-2a dataset show that EEG-TCNet achieves 77.35% classification accuracy in 4-class MI. By finding the optimal network hyperparameters per subject, we further improve the accuracy to 83.84%. Finally, we demonstrate the versatility of EEG-TCNet on the Mother of All BCI Benchmarks (MOABB), a large scale test benchmark containing 12 different EEG datasets with MI experiments. The results indicate that EEG-TCNet successfully generalizes beyond one single dataset, outperforming the current state-of-the-art (SoA) on MOABB by a meta-effect of 0.25.

Retaining persons living with HIV (PLWH) in medical care is paramount to preventing new transmissions of the virus and allowing PLWH to live normal and healthy lifespans. Maintaining regular appointments with an HIV provider and taking medication daily for a lifetime is exceedingly difficult. 51% of PLWH are non-adherent with their medications and eventually drop out of medical care. Current methods of re-linking individuals to care are reactive (after a patient has dropped-out) and hence not very effective. We describe our system to predict who is most at risk to drop-out-of-care for use by the University of Chicago HIV clinic and the Chicago Department of Public Health. Models were selected based on their predictive performance under resource constraints, stability over time, as well as fairness. Our system is applicable as a point-of-care system in a clinical setting as well as a batch prediction system to support regular interventions at the city level. Our model performs 3x better than the baseline for the clinical model and 2.3x better than baseline for the city-wide model. The code has been released on github and we hope this methodology, particularly our focus on fairness, will be adopted by other clinics and public health agencies in order to curb the HIV epidemic.

We study estimation of a gradient-sparse parameter vector $\boldsymbol{\theta}^* \in \mathbb{R}^p$, having strong gradient-sparsity $s^*:=\|\nabla_G \boldsymbol{\theta}^*\|_0$ on an underlying graph $G$. Given observations $Z_1,\ldots,Z_n$ and a smooth, convex loss function $\mathcal{L}$ for which $\boldsymbol{\theta}^*$ minimizes the population risk $\mathbb{E}[\mathcal{L}(\boldsymbol{\theta};Z_1,\ldots,Z_n)]$, we propose to estimate $\boldsymbol{\theta}^*$ by a projected gradient descent algorithm that iteratively and approximately projects gradient steps onto spaces of vectors having small gradient-sparsity over low-degree spanning trees of $G$. We show that, under suitable restricted strong convexity and smoothness assumptions for the loss, the resulting estimator achieves the squared-error risk $\frac{s^*}{n} \log (1+\frac{p}{s^*})$ up to a multiplicative constant that is independent of $G$. In contrast, previous polynomial-time algorithms have only been shown to achieve this guarantee in more specialized settings, or under additional assumptions for $G$ and/or the sparsity pattern of $\nabla_G \boldsymbol{\theta}^*$. As applications of our general framework, we apply our results to the examples of linear models and generalized linear models with random design.

This paper proposes a simple yet powerful ensemble classifier, called Random Hyperboxes, constructed from individual hyperbox-based classifiers trained on the random subsets of sample and feature spaces of the training set. We also show a generalization error bound of the proposed classifier based on the strength of the individual hyperbox-based classifiers as well as the correlation among them. The effectiveness of the proposed classifier is analyzed using a carefully selected illustrative example and compared empirically with other popular single and ensemble classifiers via 20 datasets using statistical testing methods. The experimental results confirmed that our proposed method outperformed other fuzzy min-max neural networks, popular learning algorithms, and is competitive with other ensemble methods. Finally, we identify the existing issues related to the generalization error bounds of the real datasets and inform the potential research directions.

This paper studies binary logistic regression for rare events data, or imbalanced data, where the number of events (observations in one class, often called cases) is significantly smaller than the number of nonevents (observations in the other class, often called controls). We first derive the asymptotic distribution of the maximum likelihood estimator (MLE) of the unknown parameter, which shows that the asymptotic variance convergences to zero in a rate of the inverse of the number of the events instead of the inverse of the full data sample size. This indicates that the available information in rare events data is at the scale of the number of events instead of the full data sample size. Furthermore, we prove that under-sampling a small proportion of the nonevents, the resulting under-sampled estimator may have identical asymptotic distribution to the full data MLE. This demonstrates the advantage of under-sampling nonevents for rare events data, because this procedure may significantly reduce the computation and/or data collection costs. Another common practice in analyzing rare events data is to over-sample (replicate) the events, which has a higher computational cost. We show that this procedure may even result in efficiency loss in terms of parameter estimation.

In this paper, we propose a machine learning approach for forecasting hierarchical time series. Rather than using historical or forecasted proportions, as in standard top-down approaches, we formulate the disaggregation problem as a non-linear regression problem. We propose a deep neural network that automatically learns how to distribute the top-level forecasts to the bottom level-series of the hierarchy, keeping into account the characteristics of the aggregate series and the information of the individual series. In order to evaluate the performance of the proposed method, we analyze hierarchical sales data and electricity demand data. Besides comparison with the top-down approaches, the model is compared with the bottom-up method and the optimal reconciliation method. Results demonstrate that our method does not only increase the average forecasting accuracy of the hierarchy but also addresses the need of building an automated procedure generating coherent forecasts for many time series at the same time.

Classically, ML models trained with stochastic gradient descent (SGD) are designed to minimize the average loss per example and use a distribution of training examples that remains {\em static} in the course of training. Research in recent years demonstrated, empirically and theoretically, that significant acceleration is possible by methods that dynamically adjust the training distribution in the course of training so that training is more focused on examples with higher loss. We explore {\em loss-guided training} in a new domain of node embedding methods pioneered by {\sc DeepWalk}. These methods work with implicit and large set of positive training examples that are generated using random walks on the input graph and therefore are not amenable for typical example selection methods. We propose computationally efficient methods that allow for loss-guided training in this framework. Our empirical evaluation on a rich collection of datasets shows significant acceleration over the baseline static methods, both in terms of total training performed and overall computation.

Among multiple ways of interpreting a machine learning model, measuring the importance of a set of features tied to a prediction is probably one of the most intuitive ways to explain a model. In this paper, we establish the link between a set of features to a prediction with a new evaluation criterion, robustness analysis, which measures the minimum distortion distance of adversarial perturbation. By measuring the tolerance level for an adversarial attack, we can extract a set of features that provides the most robust support for a prediction, and also can extract a set of features that contrasts the current prediction to a target class by setting a targeted adversarial attack. By applying this methodology to various prediction tasks across multiple domains, we observe the derived explanations are indeed capturing the significant feature set qualitatively and quantitatively.

Recently, federated learning (FL), as a promising distributed machine learning approach, has attracted lots of research efforts. In FL, the parameter server and the mobile devices share the training parameters over wireless links. As a result, reducing the communication overhead becomes one of the most critical challenges. Despite that there have been various communication-efficient machine learning algorithms in literature, few of the existing works consider their implementation over wireless networks. In this work, the idea of SignSGD is adopted and only the signs of the gradients are shared between the mobile devices and the parameter server. In addition, different from most of the existing works that consider Channel State Information (CSI) at both the transmitter side and the receiver side, only receiver side CSI is assumed. In such a case, an essential problem for the mobile devices is to select appropriate local processing and communication parameters. In particular, two tradeoffs are observed under a fixed total training time: (i) given the time for each communication round, the energy consumption versus the outage probability per communication round and (ii) given the energy consumption, the number of communication rounds versus the outage probability per communication round. Two optimization problems regarding the aforementioned two tradeoffs are formulated and solved. The first problem minimizes the energy consumption given the outage probability (and therefore the learning performance) requirement while the second problem optimizes the learning performance given the energy consumption requirement. Furthermore, the heterogeneous data distribution scenario is considered and a new algorithm that can deal with heterogeneous data distribution is proposed. Extensive simulations are performed to demonstrate the effectiveness of the proposed method.