Automatic detection of anomalies in space- and time-varying measurements is an important tool in several fields, e.g., fraud detection, climate analysis, or healthcare monitoring. We present an algorithm for detecting anomalous regions in multivariate spatio-temporal time-series, which allows for spotting the interesting parts in large amounts of data, including video and text data. In opposition to existing techniques for detecting isolated anomalous data points, we propose the "Maximally Divergent Intervals" (MDI) framework for unsupervised detection of coherent spatial regions and time intervals characterized by a high Kullback-Leibler divergence compared with all other data given. In this regard, we define an unbiased Kullback-Leibler divergence that allows for ranking regions of different size and show how to enable the algorithm to run on large-scale data sets in reasonable time using an interval proposal technique. Experiments on both synthetic and real data from various domains, such as climate analysis, video surveillance, and text forensics, demonstrate that our method is widely applicable and a valuable tool for finding interesting events in different types of data.

The first step in constructing a machine learning model is defining the features of the data set that can be used for optimal learning. In this work we discuss feature selection methods, which can be used to build better models, as well as achieve model interpretability. We applied these methods in the context of stress hotspot classification problem, to determine what microstructural characteristics can cause stress to build up in certain grains during uniaxial tensile deformation. The results show how some feature selection techniques are biased and demonstrate a preferred technique to get feature rankings for physical interpretations.

However, this variation is often aligned with author attributes such as age, gender, and geography, as well as more readily-available social network metadata. In this paper, we report new evidence on the link between language and social networks in the task of part-of-speech tagging. We find that tagger error rates are correlated with network structure, with high accuracy in some parts of the network, and lower accuracy elsewhere. As a result, tagger accuracy depends on training from a balanced sample of the network, rather than training on texts from a narrow subcommunity. We also describe our attempts to add robustness to stylistic variation, by building a mixture-of-experts model in which each expert is associated with a region of the social network. While prior work found that similar approaches yield performance improvements in sentiment analysis and entity linking, we were unable to obtain performance improvements in part-of-speech tagging, despite strong evidence for the link between part-of-speech error rates and social network structure.

One of the most common tasks performed by data scientists and data analysts are prediction and machine learning. This course will cover the basic components of building and applying prediction functions with an emphasis on practical applications. The course will provide basic grounding in concepts such as training and tests sets, overfitting, and error rates. The course will also introduce a range of model based and algorithmic machine learning methods including regression, classification trees, Naive Bayes, and random forests. The course will cover the complete process of building prediction functions including data collection, feature creation, algorithms, and evaluation.

A group of computer scientists have built a machine learning algorithm that can sniff out fake profiles lurking on social networks. It's likely that you and your Facebook friends have the same mutual friends. And on Twitter, it's also probable that your followers also follow the same people you do too due to common interests. These associations in social networks can be modeled on a graph as edges, where users are the vertices or nodes. The researchers from the Ben-Gurion University of the Negev, Israel, and the University of Washington, United States, hunted for fake social media accounts by developing an unsupervised learning algorithm that measures the probability of an edge existing between vertices.

An ensemble technique is characterized by the mechanism that generates the components and by the mechanism that combines them. A common way to achieve the consensus is to enable each component to equally participate in the aggregation process. A problem with this approach is that poor components are likely to negatively affect the quality of the consensus result. To address this issue, alternatives have been explored in the literature to build selective classifier and cluster ensembles, where only a subset of the components contributes to the computation of the consensus. Of the family of ensemble methods, outlier ensembles are the least studied. Only recently, the selection problem for outlier ensembles has been discussed. In this work we define a new graph-based class of ranking selection methods. A method in this class is characterized by two main steps: (1) Mapping the rankings onto a graph structure; and (2) Mining the resulting graph to identify a subset of rankings. We define a specific instance of the graph-based ranking selection class. Specifically, we map the problem of selecting ensemble components onto a mining problem in a graph. An extensive evaluation was conducted on a variety of heterogeneous data and methods. Our empirical results show that our approach outperforms state-of-the-art selective outlier ensemble techniques.

The Centers for Disease Control and Prevention (CDC) coordinates a labor-intensive process to measure the prevalence of autism spectrum disorder (ASD) among children in the United States. Random forests methods have shown promise in speeding up this process, but they lag behind human classification accuracy by about 5 percent. We explore whether newer document classification algorithms can close this gap. We applied 6 supervised learning algorithms to predict whether children meet the case definition for ASD based solely on the words in their evaluations. We compared the algorithms? performance across 10 random train-test splits of the data, and then, we combined our top 3 classifiers to estimate the Bayes error rate in the data. Across the 10 train-test cycles, the random forest, neural network, and support vector machine with Naive Bayes features (NB-SVM) each achieved slightly more than 86.5 percent mean accuracy. The Bayes error rate is estimated at approximately 12 percent meaning that the model error for even the simplest of our algorithms, the random forest, is below 2 percent. NB-SVM produced significantly more false positives than false negatives. The random forest performed as well as newer models like the NB-SVM and the neural network. NB-SVM may not be a good candidate for use in a fully-automated surveillance workflow due to increased false positives. More sophisticated algorithms, like hierarchical convolutional neural networks, would not perform substantially better due to characteristics of the data. Deep learning models performed similarly to traditional machine learning methods at predicting the clinician-assigned case status for CDC's autism surveillance system. While deep learning methods had limited benefit in this task, they may have applications in other surveillance systems.

We present a probabilistic model of an intrusion in a marked renewal process. Given a process and a sequence of events, an intrusion is a subsequence of events that is not produced by the process. Applications of the model are, for example, online payment fraud with the fraudster taking over a user's account and performing payments on the user's behalf, or unexpected equipment failures due to unintended use. We adopt Bayesian approach to infer the probability of an intrusion in a sequence of events, a MAP subsequence of events constituting the intrusion, and the marginal probability of each event in a sequence to belong to the intrusion. We evaluate the model for intrusion detection on synthetic data, as well as on anonymized data from an online payment system.

There is an increasing interest on accelerating neural networks for real-time applications. We study the student-teacher strategy, in which a small and fast student network is trained with the auxiliary information learned from a large and accurate teacher network. We propose to use conditional adversarial networks to learn the loss function to transfer knowledge from teacher to student. The proposed method is particularly effective for relatively small student networks. Moreover, experimental results show the effect of network size when the modern networks are used as student. We empirically study the trade-off between inference time and classification accuracy, and provide suggestions on choosing a proper student network.

Area under ROC (AUC) is an important metric for binary classification and bipartite ranking problems. However, it is difficult to directly optimizing AUC as a learning objective, so most existing algorithms are based on optimizing a surrogate loss to AUC. One significant drawback of these surrogate losses is that they require pairwise comparisons among training data, which leads to slow running time and increasing local storage for online learning. In this work, we describe a new surrogate loss based on a reformulation of the AUC risk, which does not require pairwise comparison but rankings of the predictions. We further show that the ranking operation can be avoided, and the learning objective obtained based on this surrogate enjoys linear complexity in time and storage. We perform experiments to demonstrate the effectiveness of the online and batch algorithms for AUC optimization based on the proposed surrogate loss.