An easy-to-hard learning paradigm for multiple classes and multiple labels

MaRDI QIDQ4637010zbMATH OpenFDO

Authors Weiwei Liu, Ivor W. Tsang, Klaus-Robert Müller

Publication date 17 April 2018

Full work available at URL http://jmlr.csail.mit.edu/papers/v18/16-212.html

zbMATH Keywords

multi-label classificationmulticlass classificationclassifier chaineasy-to-hard learning paradigm

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Learning and adaptive systems in artificial intelligence (68T05)

Recommendations

• Multiple-class classification: ordinal and categorical labels
• Empowering difficult classes with a similarity-based aggregation in multi-class classification problems
• Multi-label classification neural networks with hard logical constraints
• Training highly multiclass classifiers
• Efficient Monte Carlo methods for multi-dimensional learning with classifier chains

Cites work

• scientific article; zbMATH DE number 1149409 (Why is no real title available?)
• A simple decomposition method for support vector machines
• BoosTexter: A boosting-based system for text categorization
• Efficient distribution-free learning of probabilistic concepts
• Error-correcting tournaments
• Gaussian processes for ordinal regression
• In defense of one-vs-all classification
• LIBLINEAR: a library for large linear classification
• Learning using privileged information: SVM+ and weighted SVM
• Making decision trees feasible in ultrahigh feature and label dimensions
• Multilabel classification with principal label space transformation
• Objective-Guided Image Annotation
• Prediction and clustering in signed networks: a local to global perspective
• Scikit-learn: machine learning in Python
• Structural risk minimization over data-dependent hierarchies
• Using side information to reliably learn low-rank matrices from missing and corrupted observations

Cited in

(3)

• Learning a discriminant graph-based embedding with feature selection for image categorization
• Empowering difficult classes with a similarity-based aggregation in multi-class classification problems
• Dynamic principal projection for cost-sensitive online multi-label classification

Describes a project that uses

Uses Software

• NYU Depth
• MULAN
• Scikit

This page was built for publication: An easy-to-hard learning paradigm for multiple classes and multiple labels

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4637010)