Paris (AFP) - A computer was better than human dermatologists at detecting skin cancer in a study that pitted human against machine in the quest for better, faster diagnostics, researchers said Tuesday. A team from Germany, the United States and France taught an artificial intelligence system to distinguish dangerous skin lesions from benign ones, showing it more than 100,000 images. The machine -- a deep learning convolutional neural network or CNN -- was then tested against 58 dermatologists from 17 countries, shown photos of malignant melanomas and benign moles. Just over half the dermatologists were at "expert" level with more than five years of experience, 19 percent had between two and five years' experience, and 29 percent were beginners with less than two years under their belt. "Most dermatologists were outperformed by the CNN," the research team wrote in a paper published in the journal Annals of Oncology.

The demand of metals by modern technology has been shifting from common base metals to a variety of minor metals, such as cobalt or indium. The industrial importance and limited geological availability of some minor metals have led to them being considered more "critical," and there is a growing interest in such critical metals and their producing companies. In this research, we create a novel framework, Dynamic Advisor-Based Ensemble (dynABE), to predict the stock trend of major critical metal producers. Specifically, dynABE first utilizes domain knowledge to group the features into different "advisors," each advisor dealing with a particular economic sector. Then through ensembles of weak classifiers, each advisor produces a prediction result, and all the advisors are combined again in a biased online update fashion to dynamically make the final prediction. Based on a misclassification error of 32% for Jinchuan Group's stock (HKG: 2362), we further test a simple stock trading strategy, which leads to a back-tested return of 296%, or an excess return of 130% within one year. In addition, the feature set selected by dynABE also suggests potentially influential factors to metal criticality, because stock prices of major producers influence metal production. Therefore, not only does this research propose a novel framework for specialized stock trend prediction, it also provides domain insights into dynamic features that potentially influence metal criticality.

We present an empirical analysis of the state-of-the-art systems for referring expression recognition -- the task of identifying the object in an image referred to by a natural language expression -- with the goal of gaining insight into how these systems reason about language and vision. Surprisingly, we find strong evidence that even sophisticated and linguistically-motivated models for this task may ignore the linguistic structure, instead relying on shallow correlations introduced by unintended biases in the data selection and annotation process. For example, we show that a system trained and tested on the input image $\textit{without the input referring expression}$ can achieve a precision of 71.2% in top-2 predictions. Furthermore, a system that predicts only the object category given the input can achieve a precision of 84.2% in top-2 predictions. These surprisingly positive results for what should be deficient prediction scenarios suggest that careful analysis of what our models are learning -- and further, how our data is constructed -- is critical as we seek to make substantive progress on grounded language tasks.

Norms allow system designers to specify the desired behaviour of a sociotechnical system. In this way, norms regulate what the social and technical agents in a sociotechnical system should (not) do. In this context, a vitally important question is the development of mechanisms for monitoring whether these agents comply with norms. Proposals on norm monitoring often assume that monitoring has no costs and/or that monitors have unlimited resources to observe the environment and the actions performed by agents. In this paper, we challenge this assumption and propose the first practical resource-bounded norm monitor. Our monitor is capable of selecting the resources to be deployed and use them to check norm compliance with incomplete information about the actions performed and the state of the world. We formally demonstrate the correctness and soundness of our norm monitor and study its complexity. We also demonstrate in randomised simulations and benchmark experiments that our monitor can select monitored resources effectively and efficiently, detecting more norm violations and fulfilments than other tractable optimization approaches and obtaining slightly worse results than intractable optimal approaches.

The understanding of variations in genome sequences assists us in identifying people who are predisposed to common diseases, solving rare diseases, and finding the corresponding population group of the individuals from a larger population group. Although classical machine learning techniques allow researchers to identify groups (i.e. clusters) of related variables, the accuracy, and effectiveness of these methods diminish for large and high-dimensional datasets such as the whole human genome. On the other hand, deep neural network architectures (the core of deep learning) can better exploit large-scale datasets to build complex models. In this paper, we use the K-means clustering approach for scalable genomic data analysis aiming towards clustering genotypic variants at the population scale. Finally, we train a deep belief network (DBN) for predicting the geographic ethnicity. We used the genotype data from the 1000 Genomes Project, which covers the result of genome sequencing for 2504 individuals from 26 different ethnic origins and comprises 84 million variants. Our experimental results, with a focus on accuracy and scalability, show the effectiveness and superiority compared to the state-of-the-art.

Ensuring that all supposedly valid configurations of a software product line (SPL) lead to well-formed and acceptable products is challenging since it is most of the time impractical to enumerate and test all individual products of an SPL. Machine learning classifiers have been recently used to predict the acceptability of products associated with unseen configurations. For some configurations, a tiny change in their feature values can make them pass from acceptable to non-acceptable regarding users' requirements and vice-versa. In this paper, we introduce the idea of leveraging these specific configurations and their positions in the feature space to improve the classifier and therefore the engineering of an SPL. Starting from a variability model, we propose to use Adversarial Machine Learning techniques to create new, adversarial configurations out of already known configurations by modifying their feature values. Using an industrial video generator we show how adversarial configurations can improve not only the classifier, but also the variability model, the variability implementation, and the testing oracle.

HE FORECASTING of power demand is of crucial importance for the development of modern power systems. The stable and efficient management, scheduling and dispatch in power systems rely heavily on precise forecasting of future loads on various time horizons. In particular, shortterm load forecasting (STLF) focuses on the forecasting of loads from several minutes up to one week into the future [1]. A reliable STLF helps utilities and energy providers deal with the challenges posed by the higher penetration of renewable energies and the development of electricity markets with increasingly complex pricing strategies in future smart grids. Various STLF methods have been proposed by researchers over the years. Some of the models used for STLF include linear or nonparametric regression [2], [3], support vector regression (SVR) [1], [4], autoregressive models [5], fuzzylogic approach [6], etc. Reviews and evaluations of existing methods can be found in [7]-[10]. Building STLF systems with artificial neural networks (ANN) has long been one of the mainstream solutions to this task. As early as 2001, a review paper by Hippert et al. surveyed and examined a collection of papers that had been published between 1991 and 1999, and arrived at the conclusions that most of the proposed models

Understanding the learning dynamics of neural networks is one of the key issues for the improvement of optimization algorithms as well as for the theoretical comprehension of why deep neural nets work so well today. In this paper, we introduce a random matrix-based framework to analyze the learning dynamics of a single-layer linear network on a binary classification problem, for data of simultaneously large dimension and size, trained by gradient descent. Our results provide rich insights into common questions in neural nets, such as overfitting, early stopping and the initialization of training, thereby opening the door for future studies of more elaborate structures and models appearing in today's neural networks.

Random feature maps are ubiquitous in modern statistical machine learning, where they generalize random projections by means of powerful, yet often difficult to analyze nonlinear operators. In this paper, we leverage the "concentration" phenomenon induced by random matrix theory to perform a spectral analysis on the Gram matrix of these random feature maps, here for Gaussian mixture models of simultaneously large dimension and size. Our results are instrumental to a deeper understanding on the interplay of the nonlinearity and the statistics of the data, thereby allowing for a better tuning of random feature-based techniques.

The literature groups algorithms to learn the structure of Bayesian networks from data in three separate classes: constraint-based algorithms, which use conditional independence tests to learn the dependence structure of the data; score-based algorithms, which use goodness-of-fit scores as objective functions to maximise; and hybrid algorithms that combine both approaches. Famously, Cowell (2001) showed that algorithms in the first two classes learn the same structures when the topological ordering of the network is known and we use entropy to assess conditional independence and goodness of fit. In this paper we address the complementary question: how do these classes of algorithms perform outside of the assumptions above? We approach this question by recog-nising that structure learning is defined by the combination of a statistical criterion and an algorithm that determines how the criterion is applied to the data. Removing the confounding effect of different choices for the statistical criterion, we find using both simulated and real-world data that constraint-based algorithms do not appear to be more efficient or more sensitive to errors than score-based algorithms; and that hybrid algorithms are not faster or more accurate than constraint-based algorithms. This suggests that commonly held beliefs on structure learning in the literature are strongly influenced by the choice of particular statistical criteria rather than just properties of the algorithms themselves.