Multi-Complementary and Unlabeled Learning for Arbitrary Losses and Models
العنوان: | Multi-Complementary and Unlabeled Learning for Arbitrary Losses and Models |
---|---|
المؤلفون: | Yuzhou Cao, Yitian Xu, Shuqi Liu |
سنة النشر: | 2020 |
مصطلحات موضوعية: | FOS: Computer and information sciences, Scheme (programming language), Computer Science::Machine Learning, Computer Science - Machine Learning, Computer science, Estimator, Machine Learning (stat.ML), Sample (statistics), Unbiased Estimation, Machine Learning (cs.LG), ComputingMethodologies_PATTERNRECOGNITION, Rate of convergence, Bias of an estimator, Artificial Intelligence, Statistics - Machine Learning, Signal Processing, Learning methods, Computer Vision and Pattern Recognition, computer, Algorithm, Software, Parametric statistics, computer.programming_language |
الوصف: | A weakly-supervised learning framework named as complementary-label learning has been proposed recently, where each sample is equipped with a single complementary label that denotes one of the classes the sample does not belong to. However, the existing complementary-label learning methods cannot learn from the easily accessible unlabeled samples and samples with multiple complementary labels, which are more informative. In this paper, to remove these limitations, we propose the novel multi-complementary and unlabeled learning framework that allows unbiased estimation of classification risk from samples with any number of complementary labels and unlabeled samples, for arbitrary loss functions and models. We first give an unbiased estimator of the classification risk from samples with multiple complementary labels, and then further improve the estimator by incorporating unlabeled samples into the risk formulation. The estimation error bounds show that the proposed methods are in the optimal parametric convergence rate. Finally, the experiments on both linear and deep models show the effectiveness of our methods. 22 pages, 5 figure |
اللغة: | English |
URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_dedup___::0048763204dce30e70556ed3e8e34eb6 http://arxiv.org/abs/2001.04243 |
حقوق: | OPEN |
رقم الأكسشن: | edsair.doi.dedup.....0048763204dce30e70556ed3e8e34eb6 |
قاعدة البيانات: | OpenAIRE |
الوصف غير متاح. |