5,144 Hits in 4.0 sec

Multiple Indefinite Kernel Learning for Feature Selection

Hui Xue, Yu Song, Hai-Ming Xu
2017 Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence  
Multiple kernel learning for feature selection (MKL-FS) utilizes kernels to explore complex properties of features and performs better in embedded methods.  ...  In this paper, we propose a novel multiple indefinite kernel feature selection method (MIK-FS) based on the primal framework of indefinite kernel support vector machine (IKSVM), which applies an indefinite  ...  In the past few years, multiple kernel learning for feature selection (MKL-FS) has attracted more attention in embedded methods.  ... 
doi:10.24963/ijcai.2017/448 dblp:conf/ijcai/XueSX17 fatcat:z2mgcyayiff3nbreltgsjpxzjm

Large Scale, Large Margin Classification using Indefinite Similarity Measures [article]

Omid Aghazadeh, Stefan Carlsson
2014 arXiv   pre-print
In this paper, we investigate scalable approaches for using indefinite similarity measures in large margin frameworks.  ...  SVM using RBF kernels.  ...  Multiple Kernel Learning with PSD Kernels We tried Multiple Kernel Learning (MKL) for kernelized SVM with PSD kernels.  ... 
arXiv:1405.6922v1 fatcat:irykblt2hbgdbgrz3bf54vc75e

Scalable Learning in Reproducing Kernel Krein Spaces [article]

Dino Oglic, Thomas Gärtner
2019 arXiv   pre-print
Building on this result, we devise highly scalable methods for learning in reproducing kernel Kreĭn spaces.  ...  The devised approaches provide a principled and theoretically well-founded means to tackle large scale learning problems with indefinite kernels.  ...  Acknowledgments: We are grateful for access to the University of Nottingham High Performance Computing Facility. Dino Oglic was supported in part by EPSRC grant EP/R012067/1.  ... 
arXiv:1809.02157v2 fatcat:z7a7dfbi3nbntcd5cfblnea22u

Learning kernels from indefinite similarities

Yihua Chen, Maya R. Gupta, Benjamin Recht
2009 Proceedings of the 26th Annual International Conference on Machine Learning - ICML '09  
In order to adapt kernel methods for similarity-based learning, we introduce a method that aims to simultaneously find a reproducing kernel Hilbert space based on the given similarities and train a classifier  ...  These indefinite kernels can be problematic for standard kernel-based algorithms as the optimization problems become nonconvex and the underlying theory is invalidated.  ...  We consider it worthwhile to investigate other forms of regularizers for learning the kernel matrix from indefinite similarities.  ... 
doi:10.1145/1553374.1553393 dblp:conf/icml/ChenGR09 fatcat:lzwprwl4fraixk7dvcprwckrie

Learning sparse kernel machines with biometric similarity functions for identity recognition

Battista Biggio, Giorgio Fumera, Fabio Roli
2012 2012 IEEE Fifth International Conference on Biometrics: Theory, Applications and Systems (BTAS)  
., matching algorithms) as kernel functions.  ...  This leads us to formulate biometric recognition as a distinct two-class classification problem for each client, which can be solved even when no representation of biometric samples in a feature space  ...  to obtain a proper kernel function, or to learn kernel machines with indefinite kernels [7, 9] .  ... 
doi:10.1109/btas.2012.6374596 dblp:conf/btas/BiggioFR12 fatcat:j4sk3klg4nbdlmkwhbrrrut3xm

Indefinite Core Vector Machine

Frank-Michael Schleif, Peter Tino
2017 Pattern Recognition  
Indefinite kernels are a severe problem for most kernel based learning algorithms because classical mathematical assumptions such as positive definiteness, used in the underlying optimization frameworks  ...  set for large scale indefinite kernels with a sparse mapping.  ... 
doi:10.1016/j.patcog.2017.06.003 fatcat:msckyrunfrfrhckqoqyzrujkca

Indefinite Proximity Learning: A Review

Frank-Michael Schleif, Peter Tino
2015 Neural Computation  
We provide a comprehensive survey for the field of learning with non-metric proximities.  ...  Efficient learning of a data analysis task strongly depends on the data representation.  ...  Based on the priorly address theory multiple kernel approaches have been extended to be applicable for indefinite kernels.  ... 
doi:10.1162/neco_a_00770 pmid:26313601 fatcat:w7bp5ovxyzaovmgnncjtobwi3u

Graph Neural Networks with Composite Kernels [article]

Yufan Zhou, Jiayi Xian, Changyou Chen, Jinhui Xu
2020 arXiv   pre-print
We then propose feature aggregation as the composition of the original neighbor-based kernel and a learnable kernel to encode feature similarities in a feature space.  ...  Learning on graph structured data has drawn increasing interest in recent years.  ...  Combining K withÂ, we arrive at a composite kernel with matrixK = K Â , where denotes the element-wise multiplication. Proposition 4. MatrixK = K Â is a valid kernel matrix of an indefinite kernel.  ... 
arXiv:2005.07869v1 fatcat:eua2srwlnrepjaqjoergg7kmlq

The dissimilarity space: Bridging structural and statistical pattern recognition

Robert P.W. Duin, Elżbieta Pękalska
2012 Pattern Recognition Letters  
This leads to a set of pairwise dissimilarities from which vectors can be derived for the purpose of statistical learning.  ...  Alternatively, training sets of examples can be used in statistical learning procedures. They are most powerful for vectorial object representations.  ...  significance for integrating structural and statistical approaches to pattern recognition.  ... 
doi:10.1016/j.patrec.2011.04.019 fatcat:jfvmz75y6baehakwylvwsbhxiq

Indefinite Kernel Logistic Regression

Fanghui Liu, Xiaolin Huang, Jie Yang
2017 Proceedings of the 2017 ACM on Multimedia Conference - MM '17  
To utilize those indefinite kernels, indefinite learning methods are of great interests. This paper aims at the extension of the logistic regression from positive definite kernels to indefinite ones.  ...  Traditionally, kernel learning methods require positive definitiveness on the kernel, which is too strict and excludes many sophisticated similarities, that are indefinite.  ...  To tackle indefinite kernels in theory, the Reproducing Kernel Kreȋn Spaces (RKKS) [11] is introduced to provide a justification for feature space interpretation.  ... 
doi:10.1145/3123266.3123295 dblp:conf/mm/LiuHY17 fatcat:fvofrgswtvfovbsv3gvochoy4m

Gene ontology based transfer learning for protein subcellular localization

Suyu Mei, Wang Fei, Shuigeng Zhou
2011 BMC Bioinformatics  
For this reason, some recent predictive models are deliberately designed to integrate multiple heterogeneous data sources for exploiting multi-aspect protein feature information.  ...  The five kernels are then linearly merged into one single kernel for protein subcellular localization.  ...  Acknowledgements Thanks for the anonymous reviewers' helpful comments and the editorial office's help.  ... 
doi:10.1186/1471-2105-12-44 pmid:21284890 pmcid:PMC3039576 fatcat:pkr4elrgwnhchla6bwezzxleea

From deep to Shallow: Equivalent Forms of Deep Networks in Reproducing Kernel Krein Space and Indefinite Support Vector Machines [article]

Alistair Shilton, Sunil Gupta, Santu Rana, Svetha Venkatesh
2020 arXiv   pre-print
In this paper we explore a connection between deep networks and learning in reproducing kernel Krein space.  ...  ) kernel machine.  ...  Multiple kernel learning algorithms. Journal of Machine Learning Research, 12:2211-2268, 2011. I. S. Gradshteyn and I. M. Ryzhik. Table of Integrals, Series, and Products.  ... 
arXiv:2007.07459v2 fatcat:hg2p4fhryrap7hiccs3orrgc4m

Discriminality-driven regularization framework for indefinite kernel machine

Hui Xue, Songcan Chen
2014 Neurocomputing  
Indefinite kernel machines have attracted more and more interests in machine learning due to their better empirical classification performance than the common positive definite kernel machines in many  ...  As a result, we further present a new discriminality-driven regularization framework for indefinite kernel machine based on the discriminative regularizer.  ...  All of these methods have shown impressive improvements for the learning algorithms of indefinite kernel machines.  ... 
doi:10.1016/j.neucom.2013.11.016 fatcat:xry47ck7jvc5dhowkp46adribi

A Regularized Wasserstein Framework for Graph Kernels [article]

Asiri Wijesinghe, Qing Wang, Stephen Gould
2021 arXiv   pre-print
We propose a learning framework for graph kernels, which is theoretically grounded on regularizing optimal transport.  ...  features and their local variations, local barycenters and global connectivity.  ...  Acknowledgement: We gratefully acknowledge that the Titan Xp used for this research was donated by NVIDIA.  ... 
arXiv:2110.02554v2 fatcat:kwmv2zqiijf35o2djfbzoi7c7i

Multi-Task Learning Using Neighborhood Kernels [article]

Niloofar Yousefi, Cong Li, Mansooreh Mollaghasemi, Georgios Anagnostopoulos, Michael Georgiopoulos
2017 arXiv   pre-print
This paper introduces a new and effective algorithm for learning kernels in a Multi-Task Learning (MTL) setting.  ...  We present a Rademacher complexity bound based on which a new Multi-Task Multiple Kernel Learning (MT-MKL) model is derived.  ...  A widely adapted strategy for kernel selection is to learn a convex combination of some base kernels [15, 19] , which combined with MTL, results in the Multi-Task Multiple Kernel Learning (MT-MKL) approach  ... 
arXiv:1707.03426v1 fatcat:zx4v24y7nbcxrcetb7bgtawrvm
« Previous Showing results 1 — 15 out of 5,144 results