12,040 Hits in 5.2 sec

Unsupervised Domain Adaptation: A Multi-task Learning-based Method [article]

Jing Zhang and Wanqing Li and Philip Ogunbona
2018 arXiv   pre-print
Two novel algorithms are proposed upon the method using Regularized Least Squares and Support Vector Machines respectively.  ...  Specifically, the source and target domain classifiers are jointly learned by considering the geometry of target domain and the divergence between the source and target domains based on the concept of  ...  Regularized Least Squares Algorithm The Regularized Least Squares algorithm (denoted as mtUDA-RLS) can be expressed as, min fs,ft∈HK 1 n s ns i=1 (y i − f s (x s i )) 2 + γ I n 2 t tr(f T t Lf t ) + γ  ... 
arXiv:1803.09208v1 fatcat:cdvom35pwnhudk5wnwe35tbvcm

Unsupervised Multi-class Regularized Least-Squares Classification

Tapio Pahikkala, Antti Airola, Fabian Gieseke, Oliver Kramer
2012 2012 IEEE 12th International Conference on Data Mining  
In this work we present an efficient implementation for the unsupervised extension of the multiclass regularized least-squares classification framework, which is, to the best of the authors' knowledge,  ...  The regularized least-squares classification is one of the most promising alternatives to standard support vector machines, with the desirable property of closed-form solutions that can be obtained analytically  ...  The authors would like to thank the anonymous reviewers for valuable comments and suggestions on an early version of this work.  ... 
doi:10.1109/icdm.2012.71 dblp:conf/icdm/PahikkalaAGK12 fatcat:oefkgx6itrbsfhaycqus2rcnwe

Dictionary Learning Based on Laplacian Score in Sparse Coding [chapter]

Jin Xu, Hong Man
2011 Lecture Notes in Computer Science  
The results on binary classes datasets and multi class datasets from UCI repository demonstrate the effectiveness and efficiency of our method.  ...  Sparse coding, which is represented a vector based on sparse linear combination of a dictionary, is widely applied on signal processing, data mining and neuroscience.  ...  is based on the iteratively solving two least square optimization issue: 1 norm regularized and 2 norm constrained.  ... 
doi:10.1007/978-3-642-23199-5_19 fatcat:ryfasojsfbhl7npxsv64ndj3im

Multi-view Orthonormalized Partial Least Squares: Regularizations and Deep Extensions [article]

Li Wang and Ren-Cang Li and Wen-Wei
2020 arXiv   pre-print
Building on the least squares reformulation of OPLS, we propose a unified multi-view learning framework to learn a classifier over a common latent space shared by all views.  ...  We establish a family of subspace-based learning method for multi-view learning using the least squares as the fundamental basis.  ...  Decision values of multi-class classifiers.  ... 
arXiv:2007.05028v1 fatcat:porzpnszvbhbbgwsvuwuz3c5py

Deep Matching Autoencoders [article]

Tanmoy Mukherjee, Makoto Yamada, Timothy M. Hospedales
2017 arXiv   pre-print
We show promising results in image captioning, and on a new task that is uniquely enabled by our methodology: unsupervised classifier learning.  ...  We simultaneously optimise parameters of representation learning auto-encoders and the pairing of unpaired multi-modal data.  ...  To handle non-linearity in unsupervised multi-modal learning, kernel based approaches were proposed including Kernelized sorting (KS) [10, 26, 39] and the least-squared object matching [46, 47] .  ... 
arXiv:1711.06047v1 fatcat:cujzxw5pwbg53oyb4w2x3jje4q

Lung and Pancreatic Tumor Characterization in the Deep Learning Era: Novel Supervised and Unsupervised Learning Approaches

Sarfaraz Hussein, Pujan Kandel, Candice W. Bolan, Michael B. Wallace, Ulas Bagci
2019 IEEE Transactions on Medical Imaging  
Motivated by the radiologists' interpretations of the scans, we then show how to incorporate task-dependent feature representations into a CAD system via a graph-regularized sparse multi-task learning  ...  In the second approach, we explore an unsupervised learning algorithm to address the limited availability of labeled training data, a common problem in medical imaging applications.  ...  Thus, a least square regression function with 1 regularization can be represented as: min W XW − Y 2 2 + λ W 1 . (1) In the above equation the sparsity level of the coefficient to the bottom (attribute  ... 
doi:10.1109/tmi.2019.2894349 pmid:30676950 fatcat:woorhrucqjbcxhulknvh2irjta

Sparse, Efficient, and Semantic Mixture Invariant Training: Taming In-the-Wild Unsupervised Sound Separation [article]

Scott Wisdom, Aren Jansen, Ron J. Weiss, Hakan Erdogan, John R. Hershey
2021 arXiv   pre-print
To handle larger numbers of sources, we introduce an efficient approximation using a fast least-squares solution, projected onto the MixIT constraint set.  ...  The recent mixture invariant training (MixIT) method enables training on in-the-wild data; however, it suffers from two outstanding problems.  ...  To enable training with a large number of output sources, we proposed an efficient least-squares-based MixIT implementation.  ... 
arXiv:2106.00847v2 fatcat:spol4alhbjbypgz4ec6phf666a

Certified Robustness to Label-Flipping Attacks via Randomized Smoothing [article]

Elan Rosenfeld, Ezra Winston, Pradeep Ravikumar, J. Zico Kolter
2020 arXiv   pre-print
We generalize our results to the multi-class case, providing the first multi-class classification algorithm that is certifiably robust to label-flipping attacks.  ...  Machine learning algorithms are known to be susceptible to data poisoning attacks, where an adversary manipulates the training data to degrade performance of the resulting classifier.  ...  for pointing us to Dicker (2014) for help choosing the appropriate regularization term.  ... 
arXiv:2002.03018v4 fatcat:rm6pxwtzczhwng6n6usywzf5ti

Dual-Tree Wavelet Scattering Network with Parametric Log Transformation for Object Classification [article]

Amarjot Singh, Nick Kingsbury
2017 arXiv   pre-print
The advantages of the proposed network over other supervised and some unsupervised methods are also presented using experiments performed on different training dataset sizes.  ...  We introduce a ScatterNet that uses a parametric log transformation with Dual-Tree complex wavelets to extract translation invariant representations from a multi-resolution image.  ...  Next, orthogonal least squares (OLS) selects the subset of object class-specific dimensions across the training data, similar to that of the fully connected layers in CNNs [1] .The presence of outliers  ... 
arXiv:1702.03267v1 fatcat:qx4udnpcgbg3vnamejuz7rdwge

Discriminatively regularized least-squares classification

Hui Xue, Songcan Chen, Qiang Yang
2009 Pattern Recognition  
In this paper, we propose a novel regularization algorithm in the least-squares sense, called Discriminatively Regularized Least-Squares Classification (DRLSC) method, which is specifically designed for  ...  Furthermore, by embedding equality type constraints in the formulation, the solutions of DRLSC can follow from solving a set of linear equations and the framework naturally contains multi-class problems  ...  Acknowledgment This work was supported by National Natural Science Foundation of China (60773061).  ... 
doi:10.1016/j.patcog.2008.07.010 fatcat:2ikbssa3mvhqdmhi7cg52ebg2a

Dual Adversarial Co-Learning for Multi-Domain Text Classification [article]

Yuan Wu, Yuhong Guo
2019 arXiv   pre-print
We conduct experiments on multi-domain sentiment classification datasets. The results show the proposed approach achieves the state-of-the-art MDTC performance.  ...  under a discrepancy based co-learning framework, aiming to improve the classifiers' generalization capacity with the learned features.  ...  This method uses two forms of loss to train the domain discriminator: The negative log-likelihood loss (MAN-NLL) and the least square loss (MAN-L2). • MT-CNN: The deep neural network model which can learn  ... 
arXiv:1909.08203v1 fatcat:rrnw2zpkqbeaxlatfhpvjerhg4

Unsupervised Domain Adaptation with Regularized Domain Instance Denoising [chapter]

Gabriela Csurka, Boris Chidlowskii, Stéphane Clinchant, Sophia Michel
2016 Lecture Notes in Computer Science  
We also exploit the source class labels as another way to regularize the loss, by using a domain classifier regularizer.  ...  We report our findings and comparisons with state-of-the-art methods.  ...  In all cases, the regularization term belongs to the class of squared loss functions.  ... 
doi:10.1007/978-3-319-49409-8_37 fatcat:f4apodgxhvc5db6h2y2mnnbueq

Ensemble deep learning: A review [article]

M.A. Ganaie and Minghui Hu and A.K. Malik and M. Tanveer and P.N. Suganthan
2022 arXiv   pre-print
into ensemble models like bagging, boosting and stacking, negative correlation based deep ensemble models, explicit/implicit ensembles, homogeneous /heterogeneous ensemble, decision fusion strategies, unsupervised  ...  This paper reviews the state-of-art deep ensemble models and hence serves as an extensive summary for the researchers.  ...  The output of the models is combined via majority voting, least squares estimation weighting and double layer hierarchical approach.  ... 
arXiv:2104.02395v2 fatcat:lq73jqso5vadvnqfnnmw4zul4q

Multi-category news classification using Support Vector Machine based classifiers

Pooja Saigal, Vaibhav Khanna
2020 SN Applied Sciences  
Here, we are particularly interested in studying the behaviour of Least-squares Support Vector Machines, Twin Support Vector Machines and Least-squares Twin Support Vector Machines (LS-TWSVM) classifiers  ...  Since, all these are binary classifiers, they are extended using One-Against-All approach to handle multi-category data.  ...  Since, there is huge difference in training-testing time of least-squares classifiers and TWSVM, so y-axis is taken in logarithmic scale.  ... 
doi:10.1007/s42452-020-2266-6 fatcat:2lbdqfitsjbp5ag7xoje4nuwp4

Unsupervised and Supervised Visual Codes with Restricted Boltzmann Machines [chapter]

Hanlin Goh, Nicolas Thome, Matthieu Cord, Joo-Hwee Lim
2012 Lecture Notes in Computer Science  
Firstly, we steer the unsupervised RBM learning using a regularization scheme, which decomposes into a combined prior for the sparsity of each feature's representation as well as the selectivity for each  ...  Incorporated within the Bag of Words (BoW) framework, these techniques optimize the projection of local features into the visual codebook, leading to state-of-theart performances in many benchmark datasets  ...  Finally, we trained a linear SVM to perform multi-class classification of the test images.  ... 
doi:10.1007/978-3-642-33715-4_22 fatcat:vjoe6a7qlrdoxhlm424gnptu6y
« Previous Showing results 1 — 15 out of 12,040 results