Multilabel Classification with Principal Label Space Transformation

By Farbound Tai, Hsuan-Tien Lin. Neural Computation, 2012

329 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/8d77fb7b84a2993434c8bf3267046cecbe6bb92c

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Two-Stage Label Embedding via Neural Factorization Machine for Multi-Label Classification (2019, methodology): “The default learning rates η of all baselines (expect CPLST) and TSLE are 10 − 3 .”
Multi-label Classification via Feature-aware Implicit Label Space Encoding (2014, background): “Moreover, Tai and Lin (2010) proposed to perform principle label space transformation (PLST) for seeking important correlations between labels, which is essentially PCA (Jolliffe, 1986) for the label space.”
Adversarial Multi-label Prediction for Spoken and Visual Signal Tagging (2019, background): “Alternative label embedding approaches include principal label space dimension reduction (PLST) [5] and sparse local embedding (SLEEC) [6].”
Deep Learning for Extreme Multi-label Text Classification (2017, methodology): “These methods mainly differ in their choices of compression and decompression techniques[36] such as compressed sensing[19, 24], Bloom filters[10], Singular Value Decomposition[39], landmark labels[3], output codes[50], etc.”
Extreme Multi-label Loss Functions for Recommendation, Tagging, Ranking & Other Missing Label Applications (2016, background): “As such, the Hamming loss [4, 6, 9, 10, 13, 18, 21, 39, 42, 47] continues to be one of the most popular losses for extreme multi-label learning along with precision [3, 5, 17, 20, 21, 33, 44, 45] and the F-measure [4, 10, 14, 18, 19, 21, 39, 49].”
Sparse Local Embeddings for Extreme Multi-label Classification (2015, methodology): “Embedding methods mainly differ in the choice of their compression and decompression techniques such as compressed sensing [3], Bloom filters [4], SVD [5], landmark labels [6, 7], output codes [8], etc.”
Large Margin Metric Learning for Multi-Label Prediction (2015, methodology): “• In our experiment, PLST generally underperforms.”
Multi-label Zero-Shot Learning with Structured Knowledge Graphs (2017, background): “Label embedding methods are among the popular techniques, which transform labels into embedded label vectors, so that the correlation between labels can be exploited [47, 43, 18, 29, 49].”
Efficient Multi-label Classification with Many Labels (2013, methodology): “…proposed ML-CSSP and its kernel version ML-CSSP-Knl (with the RBF kernel) are compared with various recent multi-label output coding methods: (i) PLST (Tai & Lin, 2012); (ii) CPLST (Chen & Lin, 2012); (iii) MOPLMS (Balasubramanian & Lebanon, 2012); (iv) compressed labeling (CL) (Zhou et al., 2012).”
FastXML: a fast, accurate and stable tree-classifier for extreme multi-label learning (2014, methodology): “Methods mainly differ in the choice of compression and decompression techniques such as compressed sensing [19, 22], Bloom filters [12], SVD [29], landmark labels [4, 8], output codes [38], etc.”
Classifier chains for multi-label classification (2009, methodology): “Some other methods have also been used in the literature for reducing the complexity of binary relevance problems by taking advantage of label sparsity using compressive sensing techniques [32,16].”
Deep Extreme Multi-label Learning (2017, methodology): “erence of existing embedding based models often lies in the compression and decompression techniques for embedding and lifting respectively, e.g., compressed sensing [17], output codes [35], SVD-type [25], landmark labels [3, 6], Bloom filters [12] and etc. The recent embedding based method SLEEC [4”
On label dependence and loss minimization in multi-label classification (2012, methodology): “Simplified KDE approaches based on PCA have been studied for multi-label classification in Tai and Lin (2010).”
Large-scale Multi-label Learning with Missing Labels (2013, methodology): “The key idea in this technique is to reduce the dimensionality of the label-space by using either random projections or canonical correlation analysis (CCA) based projections Chen & Lin (2012); Hsu et al. (2009); Tai & Lin (2012); Kapoor et al. (2012).”
Multi-target regression via input space expansion: treating targets as inputs (2012, methodology): “Due to its generality, this scheme has been adopted by a number of recent methods in both MLC (Hsu 2009; Zhang and Schneider 2011, 2012; Tai and Lin 2012) and MTR (Balasubramanian and Lebanon 2012; Tsoumakas et al.”