Large Margin Metric Learning for Multi-Label Prediction

By Weiwei Liu, Ivor Wai-Hung Tsang. 2015

90 citations recorded by the source corpus.

Canonical correlation analysis (CCA) and maximum margin output coding (MMOC) methods have shown promising results for multi-label prediction, where each instance is associated with multiple labels. However, these methods require an expensive decoding procedure to recover the multiple labels of each testing instance. The testing complexity becomes unacceptable when there are many labels. To avoid decoding completely, we present a novel large margin metric learning paradigm for multi-label prediction. In particular, the proposed method learns a distance metric to discover label dependency such that instances with very different multiple labels will be moved far away. To handle many labels, we present an accelerated proximal gradient procedure to speed up the learning process. Comprehensive experiments demonstrate that our proposed method is significantly faster than CCA and MMOC in terms of both training and testing complexities. Moreover, our method achieves superior prediction performance compared with state-of-the-art methods.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Molecular Biology, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yin Cui
2. Tsung-Yi Lin
3. Serge J. Belongie
4. Jiebo Luo
5. Dacheng Tao
6. Mingli Song
7. Yale Song
8. Quoc V. Le

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Luc Van Gool
3. Jiebo Luo
4. Trevor Darrell
5. Ming-Hsuan Yang
6. Xiaogang Wang
7. Alan L. Yuille
8. Jitendra Malik

In Molecular Biology

1. Jiebo Luo
2. Yuncheng Li
3. Yale Song
4. Adriana Romero
5. Mingli Song
6. Luis Villaseñor Pineda
7. Amaia Salvador
8. Michal Drozdzal

In Evaluation & Benchmarks

1. Jiebo Luo
2. Yin Cui
3. Tsung-Yi Lin
4. Serge J. Belongie
5. Yale Song
6. Yuncheng Li
7. Mingli Song
8. Weixin Liang

Citations in the corpus

• Elucidating image-to-set prediction: An analysis of models, losses and datasets (2019, background): “Traditionally, the problem has been tackled from many different perspectives, from decomposing the problem into independent binary predictions [33, 58] or modeling label correlations [1, 43, 28], to exploiting priors such as label noise and sparsity [16, 19, 45, 54, 59, 2].”
• CU-Net: Component Unmixing Network for Textile Fiber Identification (2019, background): “2015), label dependency (Shu et al. 2015; Alessandro et al. 2013; Liu and Tsang 2015; Verma and Jawahar 2016; Zhang and Zhou 2007; Zhou et al. 2012; Wu et al. 2018), and label noise (Wu et al.”
• Two-Stage Label Embedding via Neural Factorization Machine for Multi-Label Classification (2019, background): “Thus, if we regard the factorization layer as a preprocessing procedure of data, the specialised TSLE will have the same encoding phase as a kernelized LM-kNN, where the kernel function is exactly FM.”
• DDGR: Continual Learning with Deep Diffusion-based Generative Replay (2023, background): “In the digital world, the number of new tasks is increasing overwhelmingly as huge amounts of data are produced (Liu & Tsang, 2015; Liu et al., 2017; Liu & Tsang, 2017; Liu et al., 2019).”
• Improving Pairwise Ranking for Multi-label Image Classification (2017, methodology): “[17] proposed to use largemargin metric learning by incorporating label dependency, extending the ML-kNN approach [38].”
• Collaborative Metric Learning (2017, methodology): “In this paper we apply metric learning to collaborative filtering problems and compare it to state-of-the-art collaborative filtering approaches.”