This paper presents a new learning algorithm based on decision tree ensembles. In opposition to the classical decision tree induction method, the trees of the ensemble are built by selecting the tests during their induction fully at random. This extreme

We present a novel, generic image classification method based on a recent machine learning algorithm (ensembles of extremely randomized decision trees). Images are classified using randomly extracted subwindows that are suitably normalized to yield robustness to certain image transformations. Our method is evaluated on four very different, publicly available datasets (COIL-100, ZuBuD, ETH-80, WANG). Our results show that our automatic approach is generic and robust to illumination, scale, and viewpoint changes. An extension of the method is proposed to improve its robustness with respect to rotation changes. 1.

The last decade has witnessed a tremendous growths of interests in applications that deal with querying and mining of time series data. Numerous representation methods for dimensionality reduction and similarity measures geared towards time series have been introduced. Each individual work introducing a particular method has made specific claims and, aside from the occasional theoretical justifications, provided quantitative experimental observations. However, for the most part, the comparative aspects of these experiments were too narrowly focused on demonstrating the benefits of the proposed methods over some of the previously introduced ones. In order to provide a comprehensive validation, we conducted an extensive set of time series experiments re-implementing 8 different representation methods and 9 similarity measures and their variants, and testing their effectiveness on 38 time series data sets from a wide variety of application domains. In this paper, we give an overview of these different techniques and present our comparative experimental findings regarding their effectiveness. Our experiments have provided both a unified validation of some of the existing achievements, and in some cases, suggested that certain claims in the literature may be unduly optimistic. 1.

Many algorithms have been proposed for the problem of time series classification. However, it is clear that one-nearest-neighbor with Dynamic Time Warping (DTW) distance is exceptionally difficult to beat. This approach has one weakness, however; it is computationally too demanding for many realtime applications. One way to mitigate this problem is to speed up the DTW calculations. Nonetheless, there is a limit to how much this can help. In this work, we propose an additional technique, numerosity reduction, to speed up one-nearestneighbor DTW. While the idea of numerosity reduction for nearest-neighbor classifiers has a long history, we show here that we can leverage off an original observation about the relationship between dataset size and DTW constraints to produce an extremely compact dataset with little or no loss in accuracy. We test our ideas with a comprehensive set of experiments, and show that it can efficiently produce extremely fast accurate classifiers. 1.

For many real world problems we must perform classification under widely varying amounts of computational resources. For example, if asked to classify an instance taken from a bursty stream, we may have from milliseconds to minutes to return a class prediction. For such problems an anytime algorithm may be especially useful. In this work we show how we can convert the ubiquitous nearest neighbor classifier into an anytime algorithm that can produce an instant classification, or if given the luxury of additional time, can utilize the extra time to increase classification accuracy. We demonstrate the utility of our approach with a comprehensive set of experiments on data from diverse domains.

Abstract. We present a novel approach for object category recognition that can find objects in challenging conditions using visual attention technique. It combines saliency maps very closely with the extraction of random subwindows for classification purposes. The maps are built online by the classifier while being used by it to classify the image. Saliency is therefore automatically suited to the abilities of the classifier and not an additional concept that is tried to fit into another method. Our results show that we can obtain state of the art results on commonly used datasets with using only little information and thus achieve high efficiency and short processing times. 1

Abstract. This paper presents a novel, generic, scalable, autonomous, and flexible supervised learning algorithm for the classification of multivariate and variable length time series. The essential ingredients of the algorithm are randomization, segmentation of time-series, decision tree ensemble based learning of subseries classifiers, combination of subseries classification by voting, and cross-validation based temporal resolution adaptation. Experiments are carried out with this method on 10 synthetic and real-world datasets. They highlight the good behavior of the algorithm on a large diversity of problems. Our results are also highly competitive with existing approaches from the literature. 1 Learning to classify time-series Time-series classification is an important problem from the viewpoint of its multitudinous applications. Specific applications concern the non intrusive monitoring and diagnosis of processes and biological systems, for example to decide whether the system is in a healthy operating condition on the basis of measurements

A novel and generic approach for image classification is presented. The method operates directly on pixel values and does not require feature extraction. It combines a simple local sub-window extraction technique with induction of ensembles of extremely randomized decision trees. We report results on four well known and publicly available datasets corresponding to representative applications of image classification problems: handwritten digits (MNIST), faces (ORL), 3D objects (COIL-100), and textures (OUTEX). A comparison with studies from the computer vision literature shows that our method is competitive with the state of the art, an interesting result considering its generality and conceptual simplicity. Further experiments are carried out on the COIL-100 dataset to evaluate the robustness of the learned models to rotation, scaling, or occlusion of test images. These preliminary results are very encouraging.

In this paper, we compare five tree-based machine learning methods within our recent generic image-classification framework based on random extraction and classification of subwindows. We evaluate them on three publicly available object-recognition datasets (COIL-100, ETH-80, and ZuBuD). Our comparison shows that this general and conceptually simple framework yields good results when combined with ensembles of decision trees, especially when using Tree Boosting or Extra-Trees. The latter is particularly attractive in terms of computational efficiency. 1.

Abstract- In this paper we present the basic principles of supervised learning and reinforcement learning as two complementary frameworks to design control laws or decision policies within the context of power system control. We also review recent developments in the realm of automatic learning methods and discuss their applicability to power system decision and control problems. Simulation results illustrating the potentials of the recently introduced fitted Q iteration learning algorithm in controlling a TCSC device aimed to damp electro-mechanical oscillations in a synthetic 4-machine system, are included in the paper.