In this paper, we advocate for the study of fairness techniques in low data situations. We propose two algorithms Fairness Warnings and Fair-MAML. The first is a model-agnostic algorithm that provides interpretable boundary conditions for when a fairly trained model may not behave fairly on similar but slightly different tasks within a given domain. The second is a fair meta-learning approach to train models that can be trained through gradient descent with the objective of "learning how to learn fairly". This method encodes more general notions of fairness and accuracy into the model so that it can learn new tasks within a domain both quickly and fairly from only a few training points. We demonstrate experimentally the individual utility of each model using relevant baselines for comparison and provide the first experiment to our knowledge of K-shot fairness, i.e. training a fair model on a new task with only K data points. Then, we illustrate the usefulness of both algorithms as a combined method for training models from a few data points on new tasks while using Fairness Warnings as interpretable boundary conditions under which the newly trained model may not be fair.

Description: A down-to-earth, shy but confident take on machine learning techniques that you can put to work today.

K-means -- and the celebrated Lloyd algorithm -- is more than the clustering method it was originally designed to be. It has indeed proven pivotal to help increase the speed of many machine learning and data analysis techniques such as indexing, nearest-neighbor search and prediction, data compression; its beneficial use has been shown to carry over to the acceleration of kernel machines (when using the Nystr\"om method). Here, we propose a fast extension of K-means, dubbed QuicK-means, that rests on the idea of expressing the matrix of the $K$ centroids as a product of sparse matrices, a feat made possible by recent results devoted to find approximations of matrices as a product of sparse factors. Using such a decomposition squashes the complexity of the matrix-vector product between the factorized $K \times D$ centroid matrix $\mathbf{U}$ and any vector from $\mathcal{O}(K D)$ to $\mathcal{O}(A \log A+B)$, with $A=\min (K, D)$ and $B=\max (K, D)$, where $D$ is the dimension of the training data. This drastic computational saving has a direct impact in the assignment process of a point to a cluster, meaning that it is not only tangible at prediction time, but also at training time, provided the factorization procedure is performed during Lloyd's algorithm. We precisely show that resorting to a factorization step at each iteration does not impair the convergence of the optimization scheme and that, depending on the context, it may entail a reduction of the training time. Finally, we provide discussions and numerical simulations that show the versatility of our computationally-efficient QuicK-means algorithm.

Background: While machine learning (ML) models are rapidly emerging as promising screening tools in critical care medicine, the identification of homogeneous subphenotypes within populations with heterogeneous conditions such as pediatric sepsis may facilitate attainment of high-predictive performance of these prognostic algorithms. This study is aimed to identify subphenotypes of pediatric sepsis and demonstrate the potential value of partitioned data/subtyping-based training. Methods: This was a retrospective study of clinical data extracted from medical records of 6,446 pediatric patients that were admitted at a major hospital system in the DC area. Vitals and labs associated with patients meeting the diagnostic criteria for sepsis were used to perform latent profile analysis. Modern ML algorithms were used to explore the predictive performance benefits of reduced training data heterogeneity via label profiling. Results: In total 134 (2.1%) patients met the diagnostic criteria for sepsis in this cohort and latent profile analysis identified four profiles/subphenotypes of pediatric sepsis. Profiles 1 and 3 had the lowest mortality and included pediatric patients from different age groups. Profile 2 were characterized by respiratory dysfunction; profile 4 by neurological dysfunction and highest mortality rate (22.2%). Machine learning experiments comparing the predictive performance of models derived without training data profiling against profile targeted models suggest statistically significant improved performance of prediction can be obtained. For example, area under ROC curve (AUC) obtained to predict profile 4 with 24-hour data (AUC = .998, p < .0001) compared favorably with the AUC obtained from the model considering all profiles as a single homogeneous group (AUC = .918) with 24-hour data.

In systems biology, it is common to measure biochemical entities at different levels of the same biological system. One of the central problems for the data fusion of such data sets is the heterogeneity of the data. This thesis discusses two types of heterogeneity. The first one is the type of data, such as metabolomics, proteomics and RNAseq data in genomics. These different omics data reflect the properties of the studied biological system from different perspectives. The second one is the type of scale, which indicates the measurements obtained at different scales, such as binary, ordinal, interval and ratio-scaled variables. In this thesis, we developed several statistical methods capable to fuse data sets of these two types of heterogeneity. The advantages of the proposed methods in comparison with other approaches are assessed using comprehensive simulations as well as the analysis of real biological data sets.

This paper is concerned with the factorization form of the rank regularized loss minimization problem. To cater for the scenario in which only a coarse estimation is available for the rank of the true matrix, an $\ell_{2,0}$-norm regularized term is added to the factored loss function to reduce the rank adaptively; and account for the ambiguities in the factorization, a balanced term is then introduced. For the least squares loss, under a restricted condition number assumption on the sampling operator, we establish the KL property of exponent $1/2$ of the nonsmooth factored composite function and its equivalent DC reformulations in the set of their global minimizers. We also confirm the theoretical findings by applying a proximal linearized alternating minimization method to the regularized factorizations.

We propose a framework for constructing and analyzing multiclass and multioutput classification metrics, i.e., involving multiple, possibly correlated multiclass labels. Our analysis reveals novel insights on the geometry of feasible confusion tensors -- including necessary and sufficient conditions for the equivalence between optimizing an arbitrary non-decomposable metric and learning a weighted classifier. Further, we analyze averaging methodologies commonly used to compute multioutput metrics and characterize the corresponding Bayes optimal classifiers. We show that the plug-in estimator based on this characterization is consistent and is easily implemented as a post-processing rule. Empirical results on synthetic and benchmark datasets support the theoretical findings.

Microsoft's internal big data analytics platform is comprised of hundreds of thousands of machines, serving over half a million jobs daily, from thousands of users. The majority of these jobs are recurring and are crucial for the company's operation. Although administrators spend significant effort tuning system performance, some jobs inevitably experience slowdowns, i.e., their execution time degrades over previous runs. Currently, the investigation of such slowdowns is a labor-intensive and error-prone process, which costs Microsoft significant human and machine resources, and negatively impacts several lines of businesses. In this work, we present Griffin, a system we built and have deployed in production last year to automatically discover the root cause of job slowdowns. Existing solutions either rely on labeled data (i.e., resolved incidents with labeled reasons for job slowdowns), which is in most cases non-existent or non-trivial to acquire, or on time-series analysis of individual metrics that do not target specific jobs holistically. In contrast, in Griffin we cast the problem to a corresponding regression one that predicts the runtime of a job, and show how the relative contributions of the features used to train our interpretable model can be exploited to rank the potential causes of job slowdowns. Evaluated over historical incidents, we show that Griffin discovers slowdown causes that are consistent with the ones validated by domain-expert engineers, in a fraction of the time required by them.

Accurately tracking and predicting behaviors of surrounding objects are key prerequisites for intelligent systems such as autonomous vehicles to achieve safe and high-quality decision making and motion planning. However, there still remain challenges for multi-target tracking due to object number fluctuation and occlusion. To overcome these challenges, we propose a constrained mixture sequential Monte Carlo (CMSMC) method in which a mixture representation is incorporated in the estimated posterior distribution to maintain multi-modality. Multiple targets can be tracked simultaneously within a unified framework without explicit data association between observations and tracking targets. The framework can incorporate an arbitrary prediction model as the implicit proposal distribution of the CMSMC method. An example in this paper is a learning-based model for hierarchical time-series prediction, which consists of a behavior recognition module and a state evolution module. Both modules in the proposed model are generic and flexible so as to be applied to a class of time-series prediction problems where behaviors can be separated into different levels. Finally, the proposed framework is applied to a numerical case study as well as a task of on-road vehicle tracking, behavior recognition, and prediction in highway scenarios. Instead of only focusing on forecasting trajectory of a single entity, we jointly predict continuous motions for interactive entities simultaneously. The proposed approaches are evaluated from multiple aspects, which demonstrate great potential for intelligent vehicular systems and traffic surveillance systems.

Many decision problems in science, engineering and economics are affected by uncertain parameters whose distribution is only indirectly observable through samples. The goal of data-driven decision-making is to learn a decision from finitely many training samples that will perform well on unseen test samples. This learning task is difficult even if all training and test samples are drawn from the same distribution---especially if the dimension of the uncertainty is large relative to the training sample size. Wasserstein distributionally robust optimization seeks data-driven decisions that perform well under the most adverse distribution within a certain Wasserstein distance from a nominal distribution constructed from the training samples. In this tutorial we will argue that this approach has many conceptual and computational benefits. Most prominently, the optimal decisions can often be computed by solving tractable convex optimization problems, and they enjoy rigorous out-of-sample and asymptotic consistency guarantees. We will also show that Wasserstein distributionally robust optimization has interesting ramifications for statistical learning and motivates new approaches for fundamental learning tasks such as classification, regression, maximum likelihood estimation or minimum mean square error estimation, among others.