Partial multi-label learning with noisy side information
- PDF / 2,800,997 Bytes
- 24 Pages / 439.37 x 666.142 pts Page_size
- 70 Downloads / 214 Views
Partial multi-label learning with noisy side information Lijuan Sun1,2 · Songhe Feng1,2
· Gengyu Lyu1,2 · Hua Zhang3 · Guojun Dai3
Received: 18 December 2019 / Revised: 29 October 2020 / Accepted: 1 November 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Partial multi-label learning (PML) aims to learn from the training data where each training example is annotated with a candidate label set, among which only a subset is relevant. Despite the success of existing PML approaches, a major drawback of them lies in lacking of robustness to noisy side information. To tackle this problem, we introduce a novel partial multi-label learning with noisy side information approach, which simultaneously removes noisy outliers from the training instances and trains robust partial multi-label classifier for unlabeled instances prediction. Specifically, we first represent the observed sample set as a feature matrix and then decompose it into an ideal feature matrix and an outlier feature matrix by using the low-rank and sparse decomposition scheme, where the former is constrained to be low rank by considering that the noise-free feature information always lies in a lowdimensional subspace and the latter is assumed to be sparse by considering that the outliers are usually sparse among the observed feature matrix. Secondly, we refine an ideal label confidence matrix from the observed label matrix and use the graph Laplacian regularization to constrain the confidence matrix to keep the intrinsic structure among feature vectors. Thirdly, we constrain the feature mapping matrix to be low rank by utilizing the label correlations. Finally, we obtain both the ideal features and ground-truth labels via minimizing the loss function, where the augmented Lagrange multiplier algorithm and quadratic programming are incorporated to solve the optimization problem. Extensive experiments conducted on ten different data sets demonstrate the effectiveness of our proposed method. Keywords Partial multi-label learning · Noisy side information · Low-rank and sparse decomposition
1 Introduction Partial multi-label learning (PML) is a novel multi-label learning framework, which deals with the scenario where each instance is assigned with multiple candidate labels which are only partially valid [30]. In recent years, such framework with inaccurate supervision has been widely used in many real-world applications, where accurate supervision information is difficult to be obtained from the collected data. The task of partial multi-label learning is
B
Songhe Feng [email protected]
Extended author information available on the last page of the article
123
L. Sun et al.
Fig. 1 An exemplar of partial multi-label learning with noisy side information. Among the candidate labels, rill, grass, tree, lodge, path and people are ground-truth labels, while bird, flower and mountain are irrelevant labels. Note the image features (image pixels) contain outliers. Due to the overexposure, the distant trees are blurry; this part of side inf
Data Loading...
