No abstract

Deep convolutional neural networks (CNNs) have achieved remarkable results in image processing tasks. However, their high expression ability risks overfitting. Consequently, data augmentation techniques have been proposed to prevent overfitting while enriching datasets. Recent CNN architectures with more parameters are rendering traditional data augmentation techniques insufficient. In this study, we propose a new data augmentation technique called random image cropping and patching (RICAP) which randomly crops four images and patches them to create a new training image. Moreover, RICAP mixes the class labels of the four images, resulting in an advantage of the soft labels. We evaluated RICAP with current state-ofthe-art CNNs (e.g., the shake-shake regularization model) by comparison with competitive data augmentation techniques such as cutout and mixup. RICAP achieves a new state-of-the-art test error of 2.19% on CIFAR-10. We also confirmed that deep CNNs with RICAP achieve better results on classification tasks using CIFAR-100 and ImageNet, an image-caption retrieval task using Microsoft COCO, and other computer vision tasks.

Abstract.Recently, the research on efficient extraction of previously unknown, frequently appearing patterns in a time-series data has received much attention. These patterns are called 'motifs'. Motifs are useful for various time-series data mining tasks. In this paper, we propose a motif discovery algorithm to extract a motif that represents a characteristic pattern of the given data based on Minimum Description Length (MDL) principle. In addition, the algorithm can extract motifs from multi-dimensional time-series data by using Principal Component Analysis (PCA). In experimental evaluation, we show the efficiency of the motif discovery algorithm, and the usefulness of extracted motifs to various data mining tasks.

No abstract

Over the past decade, 1 regularization has emerged as a powerful way to learn classifiers with implicit feature selection. More recently, mixed-norm (e.g., 1 / 2) regularization has been utilized as a way to select entire groups of features. In this paper, we propose a novel direct multiclass formulation specifically designed for large-scale and highdimensional problems such as document classification. Based on a multiclass extension of the squared hinge loss, our formulation employs 1 / 2 regularization so as to force weights corresponding to the same features to be zero across all classes, resulting in compact and fast-to-evaluate multiclass models. For optimization, we employ two globally-convergent variants of block coordinate descent, one with line search (