In many applications, monitoring area under the ROC curve (AUC) in a sliding window over a data stream is a natural way of detecting changes in the system. The drawback is that computing AUC in a sliding window is expensive, especially if the window size is large and the data flow is significant. In this paper we propose a scheme for maintaining an approximate AUC in a sliding window of length $k$. More specifically, we propose an algorithm that, given $\epsilon$, estimates AUC within $\epsilon / 2$, and can maintain this estimate in $O((\log k) / \epsilon)$ time, per update, as the window slides. This provides a speed-up over the exact computation of AUC, which requires $O(k)$ time, per update. The speed-up becomes more significant as the size of the window increases. Our estimate is based on grouping the data points together, and using these groups to calculate AUC. The grouping is designed carefully such that ($i$) the groups are small enough, so that the error stays small, ($ii$) the number of groups is small, so that enumerating them is not expensive, and ($iii$) the definition is flexible enough so that we can maintain the groups efficiently. Our experimental evaluation demonstrates that the average approximation error in practice is much smaller than the approximation guarantee $\epsilon / 2$, and that we can achieve significant speed-ups with only a modest sacrifice in accuracy.

Successfully predicting gentrification could have many social and commercial applications; however, real estate sales are difficult to predict because they belong to a chaotic system comprised of intrinsic and extrinsic characteristics, perceived value, and market speculation. Using New York City real estate as our subject, we combine modern techniques of data science and machine learning with traditional spatial analysis to create robust real estate prediction models for both classification and regression tasks. We compare several cutting edge machine learning algorithms across spatial, semi-spatial and non-spatial feature engineering techniques, and we empirically show that spatially-conscious machine learning models outperform non-spatial models when married with advanced prediction techniques such as feed-forward artificial neural networks and gradient boosting machine models.

Despite the progress within the last decades, weather forecasting is still a challenging and computationally expensive task. Current satellite-based approaches to predict thunderstorms are usually based on the analysis of the observed brightness temperatures in different spectral channels and emit a warning if a critical threshold is reached. Recent progress in data science however demonstrates that machine learning can be successfully applied to many research fields in science, especially in areas dealing with large datasets. We therefore present a new approach to the problem of predicting thunderstorms based on machine learning. The core idea of our work is to use the error of two-dimensional optical flow algorithms applied to images of meteorological satellites as a feature for machine learning models. We interpret that optical flow error as an indication of convection potentially leading to thunderstorms and lightning. To factor in spatial proximity we use various manual convolution steps. We also consider effects such as the time of day or the geographic location. We train different tree classifier models as well as a neural network to predict lightning within the next few hours (called nowcasting in meteorology) based on these features. In our evaluation section we compare the predictive power of the different models and the impact of different features on the classification result. Our results show a high accuracy of 96% for predictions over the next 15 minutes which slightly decreases with increasing forecast period but still remains above 83% for forecasts of up to five hours. The high false positive rate of nearly 6% however needs further investigation to allow for an operational use of our approach.

Polylingual Text Classification (PLC) consists of automatically classifying, according to a common set C of classes, documents each written in one of a set of languages L, and doing so more accurately than when naively classifying each document via its corresponding language-specific classifier. In order to obtain an increase in the classification accuracy for a given language, the system thus needs to also leverage the training examples written in the other languages. We tackle multilabel PLC via funnelling, a new ensemble learning method that we propose here. Funnelling consists of generating a two-tier classification system where all documents, irrespectively of language, are classified by the same (2nd-tier) classifier. For this classifier all documents are represented in a common, language-independent feature space consisting of the posterior probabilities generated by 1st-tier, language-dependent classifiers. This allows the classification of all test documents, of any language, to benefit from the information present in all training documents, of any language. We present substantial experiments, run on publicly available polylingual text collections, in which funnelling is shown to significantly outperform a number of state-of-the-art baselines. All code and datasets (in vector form) are made publicly available.

This paper explores two techniques to improve the performance of text-dependent speaker verification systems based on deep neural networks. Firstly, we propose a general alignment mechanism to keep the temporal structure of each phrase and obtain a supervector with the speaker and phrase information, since both are relevant for a text-dependent verification. As we show, it is possible to use different alignment techniques to replace the average pooling providing significant gains in performance. Moreover, we present a novel back-end approach to train a neural network for detection tasks by optimizing the Area Under the Curve (AUC) as an alternative to the usual triplet loss function, so the system is end-to-end, with a cost function closed to our desired measure of performance. As we can see in the experimental section, this approach improves the system performance, since our triplet AUC neural network learns how to discriminate between pairs of examples from the same identity and pairs of different identities. The different alignment techniques to produce supervectors in addition to the new back-end approach were tested on the RSR2015-Part I database for text-dependent speaker verification, providing competitive results compared to similar size networks using the average pooling to extract supervectors and using a simple back-end or triplet loss training.

To solve its task, a robot needs to have the ability to interpret its perceptions. In vision, this interpretation is particularly difficult and relies on the understanding of the structure of the scene, at least to the extent of its task and sensorimotor abilities. A robot with the ability to build and adapt this interpretation process according to its own tasks and capabilities would push away the limits of what robots can achieve in a non controlled environment. A solution is to provide the robot with processes to build such representations that are not specific to an environment or a situation. A lot of works focus on objects segmentation, recognition and manipulation. Defining an object solely on the basis of its visual appearance is challenging given the wide range of possible objects and environments. Therefore, current works make simplifying assumptions about the structure of a scene. Such assumptions reduce the adaptivity of the object extraction process to the environments in which the assumption holds. To limit such assumptions, we introduce an exploration method aimed at identifying moveable elements in a scene without considering the concept of object. By using the interactive perception framework, we aim at bootstrapping the acquisition process of a representation of the environment with a minimum of context specific assumptions. The robotic system builds a perceptual map called relevance map which indicates the moveable parts of the current scene. A classifier is trained online to predict the category of each region (moveable or non-moveable). It is also used to select a region with which to interact, with the goal of minimizing the uncertainty of the classification. A specific classifier is introduced to fit these needs: the collaborative mixture models classifier. The method is tested on a set of scenarios of increasing complexity, using both simulations and a PR2 robot.

The symptoms include proximal and axial muscle weakness, distal hyperlaxity, joint contractures, and critical respiratory insufficiency,which requires assisted ventilation and results in a reduced live expectancy. Moreover, the skin and other connective tissues where collagen VI is abundant are also affected [1, 2]. The collagen VI structural defects are related to mutations of three main genes (COL6A1, COL6A2, and COL6A3, OMIM 254090 and 158810). Thus, the new advances in genome editing tools open the possibility to successfully treat these neuromuscular diseases for the first time. This opportunity, though, comes with important challenges. Beyond the challenges of gene editing, this paper focuses on the challenges arising when trying to formally evaluate the efficiency of the therapeutic approaches in the recovery of the collagen VI microfibrillar network.

Capturing the dynamical properties of time series concisely as interpretable feature vectors can enable efficient clustering and classification for time-series applications across science and industry. Selecting an appropriate feature-based representation of time series for a given application can be achieved through systematic comparison across a comprehensive time-series feature library, such as those in the hctsa toolbox. However, this approach is computationally expensive and involves evaluating many similar features, limiting the widespread adoption of feature-based representations of time series for real-world applications. In this work, we introduce a method to infer small sets of time-series features that (i) exhibit strong classification performance across a given collection of time-series problems, and (ii) are minimally redundant. Applying our method to a set of 93 time-series classification datasets (containing over 147000 time series) and using a filtered version of the hctsa feature library (4791 features), we introduce a generically useful set of 22 CAnonical Time-series CHaracteristics, catch22. This dimensionality reduction, from 4791 to 22, is associated with an approximately 1000-fold reduction in computation time and near linear scaling with time-series length, despite an average reduction in classification accuracy of just 7%. catch22 captures a diverse and interpretable signature of time series in terms of their properties, including linear and non-linear autocorrelation, successive differences, value distributions and outliers, and fluctuation scaling properties. We provide an efficient implementation of catch22, accessible from many programming environments, that facilitates feature-based time-series analysis for scientific, industrial, financial and medical applications using a common language of interpretable time-series properties.

Better methods to detect insider threats need new anticipatory analytics to capture risky behavior prior to losing data. In search of the best overall classifier, this work empirically scores 88 machine learning algorithms in 16 major families. We extract risk features from the large CERT dataset, which blends real network behavior with individual threat narratives. We discover the predictive importance of measuring employee sentiment. Among major classifier families tested on CERT, the random forest algorithms offer the best choice, with different implementations scoring over 98% accurate. In contrast to more obscure or black-box alternatives, random forests are ensembles of many decision trees and thus offer a deep but human-readable set of detection rules (>2000 rules). We address performance rankings by penalizing long execution times against higher median accuracies using cross-fold validation. We address the relative rarity of threats as a case of low signal-to-noise (< 0.02% malicious to benign activities), and then train on both under-sampled and over-sampled data which is statistically balanced to identify nefarious actors.

AI systems excel in pattern recognition, so much so that they can stalk individual zebrafish and fruit flies even when the animals are in groups of up to a hundred. To demonstrate this, a group of researchers from the Champalimaud Foundation, a private biomedical research lab in Portugal, trained two convolutional neural networks to identify and track individual animals within a group. The aim is not so much to match or exceed humans' ability to spot and follow stuff, but rather to automate the process of studying the behavior of animals in their communities. "The ultimate goal of our team is understanding group behavior," said Gonzalo de Polavieja. "We want to understand how animals in a group decide together and learn together."