Partial label metric learning by collapsing classes
- PDF / 1,200,751 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 74 Downloads / 221 Views
ORIGINAL ARTICLE
Partial label metric learning by collapsing classes Shuang Xu1 · Min Yang1 · Yu Zhou1 · Ruirui Zheng1 · Wenpeng Liu1 · Jianjun He1 Received: 19 September 2019 / Accepted: 10 April 2020 © Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Partial label learning (PLL) is a weakly supervised learning framework proposed recently, in which the ground-truth label of training sample is not precisely annotated but concealed in a set of candidate labels, which makes the accuracy of the existing PLL algorithms is usually lower than that of the traditional supervised learning algorithms. Since the accuracy of a learning algorithm is usually closely related to its distance metric, the metric learning technologies can be employed to improve the accuracy of the existing PLL algorithms. However, only a few PLL metric learning algorithms have been proposed up to the present. In view of this, a novel PLL metric learning algorithm is proposed by using the collapsing classes model in this paper. The basic idea is first to take each training sample and its neighbor with shared candidate labels as a similar pair, while each training sample and its neighbor without shared candidate labels as a dissimilar pair, then two probability distributions are defined based on the distance and label similarity of these pairs, respectively, finally the metric matrix is obtained via minimizing the Kullback–Leibler divergence of these two probability distributions. Experimental results on six UCI data sets and four real-world PLL data sets show that the proposed algorithm can obviously improve the accuracy of the existing PLL algorithms. Keywords Partial label learning · Metric learning · Collapsing classes · Weakly supervised data
1 Introduction In the traditional supervised learning frameworks, the label information of training samples usually needs to be precisely annotated. However, in many practical applications, it is sometimes difficult to ensure that the label information of the obtained training samples is precise. Especially in the era of big data, the annotation information of data has become more diversified, which makes the traditional supervised learning technology incapable of satisfying the requirements. So weakly supervised learning frameworks such as multi-instance learning [28, 30], partial label learning [6, 33, 35], zero shot learning [14–17], and learning from crowds [21] have been widely concerned in recent years. Partial label learning (PLL) [20] is an emerging weakly supervised learning framework for handling the classification problems in which the ground-truth label of training samples is concealed in a set of candidate labels. Since PLL * Jianjun He [email protected] 1
College of Information and Communication Engineering, Dalian Minzu University, Dalian 116600, Liaoning, China
is an extension of the traditional classification framework by relaxing the constrains to training data, it also has wide application fields such as face recognition [3], part-of-speech tagging [27] and medi
Data Loading...
