Filters








91 Hits in 3.6 sec

Semi-Supervised Ordinal Regression Based on Empirical Risk Minimization [article]

Taira Tsuchiya, Nontawat Charoenphakdee, Issei Sato, Masashi Sugiyama
2021 arXiv   pre-print
To overcome these problems, we propose a novel generic framework for semi-supervised ordinal regression based on the empirical risk minimization principle that is applicable to optimizing all of the metrics  ...  Ordinal regression is aimed at predicting an ordinal class label.  ...  Supervised Ordinal Regression (SV): Supervised ordinal regression here is based on empirical risk minimization of the task surrogate risk, which is described in Pedregosa et al.  ... 
arXiv:1901.11351v3 fatcat:uhmpyssi7nbjhaltjz6vb356ta

Quadruply Stochastic Gradient Method for Large Scale Nonlinear Semi-Supervised Ordinal Regression AUC Optimization

Wanli Shi, Bin Gu, Xiang Li, Heng Huang
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this paper, we propose an unbiased objective function for S2OR AUC optimization based on ordinal binary decomposition approach.  ...  Recent researches have shown that directly optimizing concordance index or AUC can impose a better ranking on the data than optimizing the traditional error rate in ordinal regression (OR) problems.  ...  On the other hand, S 2 OR optimization problems need to handle two different types of data, i.e., unlabeled dataset and the datasets of class i, while standard DSG focuses on minimizing the empirical risk  ... 
doi:10.1609/aaai.v34i04.6029 fatcat:6btc4o373rcerkq4pkxatwjmfi

Quadruply Stochastic Gradient Method for Large Scale Nonlinear Semi-Supervised Ordinal Regression AUC Optimization [article]

Wanli Shi, Bin Gu, Xinag Li, Heng Huang
2019 arXiv   pre-print
In this paper, we propose an unbiased objective function for S^2OR AUC optimization based on ordinal binary decomposition approach.  ...  Recent researches have shown that directly optimizing concordance index or AUC can impose a better ranking on the data than optimizing the traditional error rate in ordinal regression (OR) problems.  ...  On the other hand, S 2 OR optimization problems need to handle two different types of data, i.e., unlabeled dataset and the datasets of class i, while standard DSG focuses on minimizing the empirical risk  ... 
arXiv:1912.11193v1 fatcat:vlqlib6y4jgbfktqwztnbbscty

Max-Ordinal Learning

Ines Domingues, Jaime S. Cardoso
2014 IEEE Transactions on Neural Networks and Learning Systems  
Unlike semisupervised learning, where one either has perfect knowledge about the label of the point or is completely ignorant about it, here we address a setting where, for each example, we only possess  ...  We also compare their instantiation in experiments with different base models and with conventional methods.  ...  A kernel discriminant learning ordinal regression (KDLOR) method was proposed in 2010 [5] .  ... 
doi:10.1109/tnnls.2013.2287381 fatcat:ernkhjchlngh5div4et47yauoa

RECENT ADVANCES ON SUPPORT VECTOR MACHINES RESEARCH

Yingjie Tian, Yong Shi, Xiaohui Liu
2012 Technological and Economic Development of Economy  
Second, support vector ordinal machine, semisupervised SVM, Universum SVM, robust SVM, knowledge based SVM and multi-instance SVM are then presented for nonstandard problems.  ...  Third, we explore other important issues such as lp-norm SVM for feature selection, LOOSVM based on minimizing LOO error bound, probabilistic outputs for SVM, and rule extraction from SVM.  ...  As we all know, the principal of Structural Risk Minimization (SRM) is embodied in SVM, the confidential interval and the empirical risk should be considered at the same time.  ... 
doi:10.3846/20294913.2012.661205 fatcat:vpno6pdnxjcalpsefmutch5qxq

A new approach: semi-supervised ordinal classification

2020 Turkish Journal of Electrical Engineering and Computer Sciences  
the performance of 11 our method by combining different base learners.  ...  The experiments show that the proposed method improves the classification accuracy of the model 13 compared to the existing semi-supervised method on ordinal data.  ...  [19] 2019 Empirical risk minimization Semi-supervised ordinal re- gression with Gaussian kernel (SEMI-Kernel) MAE MSE MZE Various domains [20] 2011 Max-coupled learning Semi-supervised  ... 
doi:10.3906/elk-2008-148 fatcat:aci6txlof5avpfocloi7owu3fa

Multiview Semi-supervised Learning for Ranking Multilingual Documents [chapter]

Nicolas Usunier, Massih-Reza Amini, Cyril Goutte
2011 Lecture Notes in Computer Science  
We propose a multiview learning approach to this semisupervised ranking task, where the translation of a document in a given language is considered as a view of the document.  ...  We describe a semi-supervised multiview ranking algorithm that exploits a global agreement between viewspecific ranking functions on a set of unlabeled observations.  ...  A direct extension of our work is to examine the possibility of multiview, semisupervised ranking when the reference ranking information is not bipartite, but take the form of either scores on an ordinal  ... 
doi:10.1007/978-3-642-23808-6_29 fatcat:xdmqogqynje2tjfsdvzrfj4p7m

A Novel Variational Autoencoder with Applications to Generative Modelling, Classification, and Ordinal Regression [article]

Joel Jaskari, Jyri J. Kivinen
2018 arXiv   pre-print
We develop a novel probabilistic generative model based on the variational autoencoder approach.  ...  We describe how to do supervised, unsupervised and semi-supervised learning, and nominal and ordinal classification, with the model.  ...  There has been some previous work on semi-supervised ordinal regression.  ... 
arXiv:1812.07352v2 fatcat:yvk5gcnlyzadfl23sgmjxgmx2q

Regression with Comparisons: Escaping the Curse of Dimensionality with Ordinal Information [article]

Yichong Xu, Sivaraman Balakrishnan, Aarti Singh, Artur Dubrawski
2019 arXiv   pre-print
In this paper, we consider a semi-supervised regression setting, where we obtain additional ordinal (or comparison) information for the unlabeled samples.  ...  We provide a precise quantification of the usefulness of these types of ordinal feedback in both nonparametric and linear regression, showing that in many cases it is possible to accurately estimate an  ...  Acknowledgements We thank Hariank Muthakana for his help on the age prediction experiments.  ... 
arXiv:1806.03286v2 fatcat:lqxz32drsbapnmvlzu4efzdscu

Robust regression on image manifolds for ordered label denoising

Hui Wu, Richard Souvenir
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
In this paper, we present a computationally efficient and non-parametric method for robust regression on manifolds.  ...  ., real-valued, ordinal) labels.  ...  One such approach is based on the Hessian regularizer, which has been applied to, for example, nonlinear dimensionality reduction [6] and semisupervised regression [6] .  ... 
doi:10.1109/cvpr.2015.7298627 dblp:conf/cvpr/WuS15 fatcat:bagjoe2syzey5kfdo2kfgt5s5y

An RKHS for multi-view learning and manifold co-regularization

Vikas Sindhwani, David S. Rosenberg
2008 Proceedings of the 25th international conference on Machine learning - ICML '08  
unlabeled examples, and (b) the average prediction given by the chosen functions performs well on labeled examples.  ...  semi-supervised kernel methods implement the following idea: find a function in each of multiple Reproducing Kernel Hilbert Spaces (RKHSs) such that (a) the chosen functions make similar predictions on  ...  Note that while Theorem 3.4 bounds the gap between expected and empirical performance of an arbitrary f ∈H 1 , Theorem 3.5 bounds the gap between the empirical loss minimizer overH 1 and true risk minimizer  ... 
doi:10.1145/1390156.1390279 dblp:conf/icml/SindhwaniR08 fatcat:s3erisq2lff2zcbbwvoln6wpsm

Learning to rank for information retrieval

Tie-Yan Liu
2010 Proceeding of the 33rd international ACM SIGIR conference on Research and development in information retrieval - SIGIR '10  
The book is completed by theoretical discussions on guarantees for ranking performance, and the outlook of future research on learning to rank.  ...  This book also provides several promising future research v vi Preface directions on learning to rank, hoping that the readers can be inspired to work on these new topics and contribute to this emerging  ...  Based on the pairwise 0-1 loss, the expected risks and empirical risks are defined as below. <0} .  ... 
doi:10.1145/1835449.1835676 dblp:conf/sigir/Liu10 fatcat:ud536unb5ng7pcijqtnlmrxp6y

Synthetic semi-supervised learning in imbalanced domains: Constructing a model for donor-recipient matching in liver transplantation

M. Pérez-Ortiz, P.A. Gutiérrez, M.D. Ayllón-Terán, N. Heaton, R. Ciria, J. Briceño, C. Hervás-Martínez
2017 Knowledge-Based Systems  
Dutkowski et al. (2011) recently proposed a balance of risk (BAR) score based on donor and recipient characteristics.  ...  Semisupervised learning has being studied mainly for binary classification (Cai et al., 2007; Cohen et al., 2004) and regression (Zhu, 2005) .  ... 
doi:10.1016/j.knosys.2017.02.020 fatcat:pklgs5rlonbonpy5gh4zzbswua

Toward the Automatic Updating of Land-Cover Maps by a Domain-Adaptation SVM Classifier and a Circular Validation Strategy

L. Bruzzone, M. Marconcini
2009 IEEE Transactions on Geoscience and Remote Sensing  
The problem is modeled in the domain-adaptation framework by introducing a novel method designed for land-cover map updating, which is based on a domain-adaptation support vector machine technique.  ...  Experimental results obtained on a multitemporal and multispectral data set confirmed the effectiveness and the reliability of the proposed system.  ...  In order to increase the reliability of the semisupervised learning process, systems based on ensemble methods have also been devised.  ... 
doi:10.1109/tgrs.2008.2007741 fatcat:lzreq7dzxvfvnpemwm7st66gci

Error Analysis of Stochastic Gradient Descent Ranking

Hong Chen, Yi Tang, Luoqing Li, Yuan Yuan, Xuelong Li, Yuanyan Tang
2013 IEEE Transactions on Cybernetics  
A kernel-based stochastic gradient descent algorithm with the least squares loss is proposed for ranking in this paper.  ...  Experimental results on real-world data have shown the effectiveness of the proposed algorithm in ranking tasks, which verifies the theoretical analysis in ranking error.  ...  Methods Based on Pairwise Relations Some of the earlier studies, such as [19] , [27] , and [28] , focus on the problem of ordinal regression. Freund et al.  ... 
doi:10.1109/tsmcb.2012.2217957 pmid:24083315 fatcat:y5bvq4tjwrgw3iamjthy5euz5a
« Previous Showing results 1 — 15 out of 91 results