Filters








166 Hits in 3.7 sec

A regularization framework for multiclass classification: A deterministic annealing approach

Zhihua Zhang, Gang Wang, Dit-Yan Yeung, Guang Dai, Frederick Lochovsky
2010 Pattern Recognition  
In this paper, we propose a general regularization framework for multiclass classification based on discriminant functions.  ...  With the aid of the deterministic annealing approach, a differentiable objective function is derived subject to a constraint on the randomness of the solution.  ...  By applying deterministic annealing to this regularization framework for multiclass classification, we obtain an algorithm called annealed discriminant analysis (ADA).  ... 
doi:10.1016/j.patcog.2010.02.003 fatcat:z2n6dkhiejhzvmsnjodmj6sfn4

Online Deterministic Annealing for Classification and Clustering [article]

Christos Mavridis, John Baras
2021 arXiv   pre-print
We introduce an online prototype-based learning algorithm that can be viewed as a progressively growing competitive-learning neural network architecture for classification and clustering.  ...  an annealing process.  ...  Online Deterministic Annealing for Classification We can extend the proposed learning algorithm to be used for classification as well.  ... 
arXiv:2102.05836v3 fatcat:xfpsmhipzrcepcv2ljdg7gjcqa

Deterministic annealing for clustering, compression, classification, regression, and related optimization problems

K. Rose
1998 Proceedings of the IEEE  
The deterministic annealing approach to clustering and its extensions has demonstrated substantial performance improvement over standard supervised and unsupervised learning methods in a variety of important  ...  It is derived within a probabilistic framework from basic information theoretic principles (e.g., maximum entropy and random coding).  ...  DETERMINISTIC ANNEALING FOR UNSUPERVISED LEARNING A.  ... 
doi:10.1109/5.726788 fatcat:zwthlhtyr5ffpkvavsojkcq5ny

Multicategory large margin classification methods: Hinge losses vs. coherence functions

Zhihua Zhang, Cheng Chen, Guang Dai, Wu-Jun Li, Dit-Yan Yeung
2014 Artificial Intelligence  
Finally, we develop multicategory large margin classification methods by using a so-called multiclass C-loss.  ...  In particular, we explore the Fisher consistency properties of multicategory majorization losses and present a construction framework of majorization losses of the 0-1 loss.  ...  Acknowledgements The authors would like to thank the three anonymous referees for their insightful comments on the original version of this paper. Wu  ... 
doi:10.1016/j.artint.2014.06.002 fatcat:6psadlz6qjbi3koxkfhmphnqye

Multiclass Classification with Multi-Prototype Support Vector Machines

Fabio Aiolli, Alessandro Sperduti
2005 Journal of machine learning research  
For this problem, we give a compact constrained quadratic formulation and we propose a greedy optimization algorithm able to find locally optimal solutions for the non convex objective function.  ...  Winner-take-all multiclass classifiers are built on the top of a set of prototypes each representing one of the available classes.  ...  Crammer, and the anonymous reviewers for their valuable suggestions on how to improve the paper.  ... 
dblp:journals/jmlr/AiolliS05 fatcat:novu6e6e4bhgfcnwcrsmcy2b2a

From categories to subcategories: Large-scale image classification with partial class label refinement

Marko Ristin, Juergen Gall, Matthieu Guillaumin, Luc Van Gool
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
To this end, we adopt the framework of Random Forests and propose a regularized objective function that takes into account relations between categories and subcategories.  ...  Compared to approaches that disregard the extra coarse labeled data, we achieve a relative improvement in subcategory classification accuracy of up to 22% in our large-scale image classification experiments  ...  The authors acknowledge financial support from the CTI project (15769.1 PFES-ES), DFG Emmy Noether program (GA 1927/1-1), DFG project (GA 1927/2-2 FOR 1505) and Toyota.  ... 
doi:10.1109/cvpr.2015.7298619 dblp:conf/cvpr/RistinGGG15 fatcat:3dvgmtzsvnainkzy3djitnx5w4

Evidential Relational-Graph Convolutional Networks for Entity Classification in Knowledge Graphs

Tobias Weller, Heiko Paulheim
2021 Proceedings of the 30th ACM International Conference on Information & Knowledge Management  
In addition, the experiments show that this approach overcomes the well-known problem of overconfident prediction of deterministic neural networks.  ...  We use the continuous output of a Graph Convolutional Neural Network as parameters for a Dirichlet distribution.  ...  Furthermore, we would like to apply this approach within an active learning framework, using the uncertainty metric for determining which data points should be labeled.  ... 
doi:10.1145/3459637.3482102 fatcat:h5xpav7i2bhdpkgweri3yd57ie

Efficient recurrent local search strategies for semi- and unsupervised regularized least-squares classification

Fabian Gieseke, Oliver Kramer, Antti Airola, Tapio Pahikkala
2012 Evolutionary Intelligence  
In general, sufficient labeled data is needed for such classification settings to obtain reasonable models.  ...  We evalu-ate the performances of the resulting approaches on a variety of artificial and real-world data sets. The results indicate that our approaches can successfully incorporate unlabeled data. 1  ...  The Cholesky decomposition for a m × m-matrix can be obtained in O(m 3 ) time (in practice and up to machine precision, see Golub and Van Loan (1989) pp. 141-145).  ... 
doi:10.1007/s12065-012-0068-5 fatcat:ahgubi7cjbcl7bhvtxzquyldgi

Deformed Kernel Based Extreme Learning Machine

Chen Zhang, Xiong Shi Xia, Bing Liu
2013 Journal of Computers  
The extreme learning machine (ELM) is a newly emerging supervised learning method.  ...  The experimental results showed that the proposed semi-supervised extreme learning machine tends to achieve outstanding generalization performance at a relatively faster learning speed than traditional  ...  For L 2 - 2 TSVM-MFN, DA L 2 -SVM-MFN and CutS 3 VM, multiclass datasets are learned using a one-versus-rest approach.  ... 
doi:10.4304/jcp.8.6.1602-1609 fatcat:kwpr3yvtvvbajmv3tdnmfn3ace

Adaptive Feature Spaces for Land Cover Classification with Limited Ground Truth Data [chapter]

Joseph T. Morgan, Alex Henneguelle, Melba M. Crawford, Joydeep Ghosh, Amy Neuenschwander
2002 Lecture Notes in Computer Science  
using a variety of direct approaches to the multiclass problem.  ...  Classical techniques for dealing with small sample sizes include regularization of covariance matrices and feature reduction.  ...  We thank Amy Neuenschander and Yangchi Chen for their help with data preparation and interpretation of results.  ... 
doi:10.1007/3-540-45428-4_19 fatcat:it5mij6ajfcfjfy2beykucisny

ADAPTIVE FEATURE SPACES FOR LAND COVER CLASSIFICATION WITH LIMITED GROUND TRUTH DATA

JOSEPH T. MORGAN, JISOO HAM, MELBA M. CRAWFORD, ALEX HENNEGUELLE, JOYDEEP GHOSH
2004 International journal of pattern recognition and artificial intelligence  
using a variety of direct approaches to the multiclass problem.  ...  Classical techniques for dealing with small sample sizes include regularization of covariance matrices and feature reduction.  ...  We thank Amy Neuenschander and Yangchi Chen for their help with data preparation and interpretation of results.  ... 
doi:10.1142/s0218001404003411 fatcat:fo4wrvamlba53o77gb2njxvvye

Large Margin Semi-supervised Structured Output Learning [article]

P. Balamurugan, Shirish Shevade, Sundararajan Sellamanickam
2013 arXiv   pre-print
The alternating optimization is carried out within a deterministic annealing framework, which helps in effective constraint matching, and avoiding local minima which are not very useful.  ...  We propose a simple optimization approach, which alternates between solving a supervised learning problem and a constraint matching problem.  ...  Hence, the deterministic annealing framework provides a useful way to get a reasonable value of C u .  ... 
arXiv:1311.2139v1 fatcat:4gerfx7b5jbqvh62gdb2sqob2m

Modeling a deep transfer learning framework for the classification of COVID-19 radiology dataset

Michael Adebisi Fayemiwo, Toluwase Ayobami Olowookere, Samson Afolabi Arekete, Adewale Opeoluwa Ogunde, Mba Obasi Odim, Bosede Oyenike Oguntunde, Oluwabunmi Omobolanle Olaniyan, Theresa Omolayo Ojewumi, Idowu Sunday Oyetade, Ademola Adegoke Aremu, Aderonke Anthonia Kayode
2021 PeerJ Computer Science  
Moreover, the fine-tuned VGG-16 and VGG-19 models have MCC of 0.98 and 0.96 respectively in the binary classification, and 0.91 and 0.89 for multiclass classification.  ...  The VGG-16 DTL model has a Kappa value of 0.98 as against 0.96 for the VGG-19 DTL model in the binary classification task, while in the three-class classification problem, the VGG-16 DTL model has a Kappa  ...  ADDITIONAL INFORMATION AND DECLARATIONS Funding The authors received no funding for this work. Competing Interests The authors declare that they have no competing interests.  ... 
doi:10.7717/peerj-cs.614 fatcat:ralfizgflzdc7gjtupcpga2qxm

A New Cascade Model for the Hierarchical Joint Classification of Multitemporal and Multiresolution Remote Sensing Data

Ihsen Hedhli, Gabriele Moser, Josiane Zerubia, Sebastiano Bruno Serpico
2016 IEEE Transactions on Geoscience and Remote Sensing  
In this paper, we propose a novel method for the joint classification of both multidate and multiresolution remote sensing imagery, which represents an important and relatively unexplored classification  ...  Within this framework, a novel element of the proposed approach is the use of multiple quad-trees in cascade, each associated with the images available at each observation date in the considered time series  ...  In other words, it is often a good tradeoff between a deterministic gradient-like algorithm and the simulated annealing algorithm.  ... 
doi:10.1109/tgrs.2016.2580321 fatcat:awgh7f33avbrfhxqsiqfrloqmu

Semi-Supervised Random Forests

Christian Leistner, Amir Saffari, Jakob Santner, Horst Bischof
2009 2009 IEEE 12th International Conference on Computer Vision  
From this intuition, we develop a novel multi-class margin definition for the unlabeled data, and an iterative deterministic annealing-style training algorithm maximizing both the multi-class margin of  ...  Furthermore, we propose a control mechanism based on the out-of-bag error, which prevents the algorithm from degradation if the unlabeled data is not useful for the task.  ...  Based on that, we propose to optimize a regularized loss function with the usage of Deterministic Annealing. Margin for Multi-Class Classification Recently, Zou et al.  ... 
doi:10.1109/iccv.2009.5459198 dblp:conf/iccv/LeistnerSSB09 fatcat:rfziptxqbjcothokvnyp7vouwm
« Previous Showing results 1 — 15 out of 166 results