Department of Computer Science, Rutgers University, CBIM, Piscataway Township, NJ 08854, USA.
Department of Pathology, Immunology and Laboratory Medicine, Rutgers University- New Jersey Medical School, 185 South Orange Avenue, Newark, NJ, 07103, USA; Center for Immunity and Inflammation, New Jersey Medical School, Rutgers-The State University of New Jersey, Newark, NJ, 07103, USA.
Neural Netw. 2021 Mar;135:68-77. doi: 10.1016/j.neunet.2020.11.015. Epub 2020 Dec 13.
The goal of n-shot learning is the classification of input data from small datasets. This type of learning is challenging in neural networks, which typically need a high number of data during the training process. Recent advancements in data augmentation allow us to produce an infinite number of target conditions from the primary condition. This process includes two main steps for finding the best augmentations and training the data with the new augmentation techniques. Optimizing these two steps for n-shot learning is still an open problem. In this paper, we propose a new method for auto-augmentation to address both of these problems. The proposed method can potentially extract many possible types of information from a small number of available data points in n-shot learning. The results of our experiments on five prominent n-shot learning datasets show the effectiveness of the proposed method.
小样本学习的目标是对小数据集的输入数据进行分类。在神经网络中，这种学习类型具有挑战性，因为神经网络通常在训练过程中需要大量数据。最近的数据增强技术的进步允许我们从主要条件中生成无限数量的目标条件。这个过程包括寻找最佳增强和使用新的增强技术训练数据的两个主要步骤。针对小样本学习优化这两个步骤仍然是一个开放的问题。在本文中，我们提出了一种新的自动增强方法来解决这两个问题。该方法可以从小样本学习中可用的少量数据点中提取许多可能的信息类型。我们在五个著名的小样本学习数据集上的实验结果表明了该方法的有效性。
