Multi-Label Learning with Global and Local Label Correlation

Zhu, Yue; Kwok, James T.; Zhou, Zhi‐Hua

doi:10.1109/tkde.2017.2785795

Cited by 319 publications

(113 citation statements)

References 35 publications

Supporting

Mentioning

111

Contrasting

Order By: Relevance

“…Another line of related works is multi-label learning with missing labels. Many approaches aim to recover missing labels by exploiting low rank structure or label correlations [1], [12], [29], [35]. In their settings, the labels, which are missing in some instances, are within the observed class label set.…”

Section: Related Workmentioning

confidence: 99%

Multi-Label Learning with Emerging New Labels

Zhu

Ting

Zhou

2018

IEEE Trans. Knowl. Data Eng.

Self Cite

View full text Add to dashboard Cite

In a multi-label learning task, an object possesses multiple concepts where each concept is represented by a class label. Previous studies on multi-label learning have focused on a fixed set of class labels, i.e., the class label set of test data is the same as that in the training set. In many applications, however, the environment is dynamic and new concepts may emerge in a data stream. In order to maintain a good predictive performance in this environment, a multi-label learning method must have the ability to detect and classify instances with emerging new labels. To this end, we propose a new approach called Multi-label learning with Emerging New Labels (MuENL). It has three functions: classify instances on currently known labels, detect the emergence of a new label, and construct a new classifier for each new label that works collaboratively with the classifier for known labels. In addition, we show that MuENL can be easily extended to handle sparse high dimensional data streams by simply reducing the original dimensionality, and then applying MuENL on the reduced dimensional space. Our empirical evaluation shows the effectiveness of MuENL on several benchmark datasets and MuENLHD on the sparse high dimensional Weibo dataset.

show abstract

Section: Related Workmentioning

confidence: 99%

Multi-Label Learning with Emerging New Labels

Zhu

Ting

Zhou

2018

IEEE Trans. Knowl. Data Eng.

Self Cite

View full text Add to dashboard Cite

show abstract

“…We compare our approach to the state-of-the-art emotion tagging methods: MET [20], TRBM [19], FRBM [23], and CEDM [18]. We also compare it to several multi-label classification methods, i.e., GLOCAL [27] and CLP-RNN [11]. Experimental results are shown in Table 2.…”

Section: Comparisons To Related Workmentioning

confidence: 99%

Exploiting Multi-Emotion Relations at Feature and Label Levels for Emotion Tagging

Wang

2020

Proceedings of the 28th ACM International Conference on Multimedia

View full text Add to dashboard Cite

The dependence among emotions is crucial to boost emotion tagging. In this paper, we propose a novel emotion tagging method, that thoroughly explores emotion relations from both the feature and label levels. Specifically, a graph convolutional network is introduced to inject local dependence among emotions into the model at the feature level, while an adversarial learning strategy is applied to constrain the joint distribution of multiple emotions at the label level. In addition, a new balanced loss function that mitigates the adverse effects of intra-class and inter-class imbalance is introduced to deal with the imbalance of emotion labels. Experimental results on several benchmark databases demonstrate the superiority of the proposed method compared to state-of-the-art works. CCS CONCEPTS • Human-centered computing → HCI design and evaluation methods; Social media.

show abstract

“…Inspired by [20], we separate the whole training data ] ∈ R l×n m be the corresponding label output matrix, where f m,j,: ∈ R 1×n (j = 1, 2, . .…”

Section: Incorporating Local Label Correlationsmentioning

confidence: 99%

“…To validate the efficacy of the proposed framework, we conduct comparisons on nine common-used benchmark data sets, 1 which have been widely used in MLFS literature [15], [19], [20], [25], [26] To evaluate the performance of comparing methods, we employ two types of evaluation metrics in MLL, i.e., example-based and label-based [1]. Given a test data set…”

Section: Experimental Study a Data Setsmentioning

confidence: 99%

“…To alleviate this, a MLL method, called ML-LOC, exploits label correlations locally, which extends the feature representation of each instance by embedding a code that encodes the influence of the instance's labels to the local label correlations [19]. Though several works on MLL have been designed to learn and exploit local label correlations [20], [21], it has not been considered in MLFS. Thus, our framework tries to incorporate the exploitation of local label correlations into feature learning.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Exploring Common and Label-Specific Features for Multi-Label Learning With Local Label Correlations

et al. 2020

View full text Add to dashboard Cite

In multi-label learning, instances can be associated with a set of class labels. The existing multilabel feature selection (MLFS) methods generally adopt either of these two strategies, namely, selecting a subset of features that is shared by all labels (common features) or exploring the most discriminative features for each label (label-specific features). However, both of them can play a key role in the discrimination of different labels. For example, common features can distinguish all labels, and label-specific features contribute to discriminating label's differences. They are important for the discriminability of selected features. On the other hand, it is well-known that exploiting label correlations can advance the performance of MLFS, and label correlations are local and only shared by a data subset in most cases. How to effectively learn and exploit local label correlations in the selection process is significant. In this paper, to address these problems, we propose a novel MLFS framework. Specially, common and label-specific features are simultaneously considered by introducing both l 2,1-norm and l 1-norm regularizers, local label correlations are automatically learned with probability and learned correlation information is efficiently exploited to help feature selection by constraining label correlations on the output of labels. A comparative study with seven state-of-the-art methods manifests the efficacy of our framework.

show abstract

Multi-Label Learning with Global and Local Label Correlation

Cited by 319 publications

References 35 publications

Multi-Label Learning with Emerging New Labels

Multi-Label Learning with Emerging New Labels

Exploiting Multi-Emotion Relations at Feature and Label Levels for Emotion Tagging

Exploring Common and Label-Specific Features for Multi-Label Learning With Local Label Correlations

Contact Info

Product

Resources

About