434 Hits in 5.6 sec

Learning and Selecting Features Jointly with Point-wise Gated Boltzmann Machines

Kihyuk Sohn, Guanyu Zhou, Chansoo Lee, Honglak Lee
2013 International Conference on Machine Learning  
To address this problem, we propose a point-wise gated Boltzmann machine, a unified generative model that combines feature learning and feature selection.  ...  Our model performs not only feature selection on learned high-level features (i.e., hidden units), but also dynamic feature selection on raw features (i.e., visible units) through a gating mechanism.  ...  Acknowledgments This work was supported in part by NSF IIS 1247414 and a Google Faculty Research Award.  ... 
dblp:conf/icml/SohnZLL13 fatcat:4xcsch3op5ahjndycwqr5bsvxu

Jointly Feature Learning and Selection for Robust Tracking via a Gating Mechanism

Bineng Zhong, Jun Zhang, Pengfei Wang, Jixiang Du, Duansheng Chen, Quan Zou
2016 PLoS ONE  
To address this problem, we propose a novel visual tracking method via a point-wise gated convolutional deep network (CPGDN) that jointly performs the feature learning and feature selection in a unified  ...  The proposed method performs dynamic feature selection on raw features through a gating mechanism.  ...  We construct a two-layer CPGDN model, in which the first layer is composed by convolutional restricted Boltzmann machines (CRBM) and the second layer is composed by convolutional point-wise gated Boltzmanne  ... 
doi:10.1371/journal.pone.0161808 pmid:27575684 pmcid:PMC5004979 fatcat:oizomflejrbt5ajei5aej4kbpe

Learning to Disentangle Factors of Variation with Manifold Interaction

Scott E. Reed, Kihyuk Sohn, Yuting Zhang, Honglak Lee
2014 International Conference on Machine Learning  
To address this, we propose a higher-order Boltzmann machine that incorporates multiplicative interactions among groups of hidden units that each learn to encode a distinct factor of variation.  ...  Many existing feature learning algorithms focus on a single task and extract features that are sensitive to the task-relevant factors and invariant to all others.  ...  DGE 1256260, and the Google Faculty Research Award.  ... 
dblp:conf/icml/ReedSZL14 fatcat:qg7nhbhiajghfcmw5namwtgnpm

Intrusion Detection using Machine Learning and Deep Learning

2019 International journal of recent technology and engineering  
A number of techniques came into existence to detect the intrusions on the basis of machine learning and deep learning procedures.  ...  With the increase in usage of networking technology and the Internet, Intrusion detection becomes important and challenging security problem.  ...  Deep Belief Networks (DBN): DBN is a store of limited Boltzmann machines, in which each RBM layer talks with both the past and coming about layers.  ... 
doi:10.35940/ijrte.d9999.118419 fatcat:nc5iw6tmbzh7pbtzuzcfhtxoxi


2017 International Journal of Recent Trends in Engineering and Research  
Existing methods proposed to use Canonical Correlation Analysis (CCA), manifolds learning, dual-wing harmoniums, deep autoencoder, and deep Boltzmann machine to approach the task.  ...  Images and videos are associated with tags and captions. According to research published on eMarketer, about 75 percent of the content posted by Facebook users contains photos.  ...  Taylor et al. proposed a convolutional gated restricted Boltzmann machineto model the spatio-temporal features for videos. @IJRTER-2017, All Rights Reserved  ... 
doi:10.23883/ijrter.2017.3365.aeikk fatcat:6dmfmfsmtbaejale6t63ts7may

Robust Deep Appearance Models [article]

Kha Gia Quach, Chi Nhan Duong, Khoa Luu, Tien D. Bui
2016 arXiv   pre-print
The two models are connected by Restricted Boltzmann Machines at the top layer to jointly learn and capture the variations of both facial shapes and appearances.  ...  In this approach, two crucial components of face images, i.e. shape and texture, are represented by Deep Boltzmann Machines and Robust Deep Boltzmann Machines (RDBM), respectively.  ...  ACKNOWLEDGMENT This work is supported in part by the Natural Sciences and Engineering Research Council (NSERC) of Canada  ... 
arXiv:1607.00659v1 fatcat:hisxvnqt7rebbf5pb2ap7dnvaa

Deep Learning and Its Applications to Machine Health Monitoring: A Survey [article]

Rui Zhao, Ruqiang Yan, Zhenghua Chen, Kezhi Mao, Peng Wang, Robert X. Gao
2016 arXiv   pre-print
variants, Restricted Boltzmann Machines and its variants including Deep Belief Network (DBN) and Deep Boltzmann Machines (DBM), Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN).  ...  The main purpose of this paper is to review and summarize the emerging research work of deep learning on machine health monitoring.  ...  At last, the above three modules including feature design, feature extraction/selection and model training can not be jointly optimized which may hinder the final performance of the whole system.  ... 
arXiv:1612.07640v1 fatcat:46zurdmo2bcabo352qqppv5zki

Learning to Relate Images

R. Memisevic
2013 IEEE Transactions on Pattern Analysis and Machine Intelligence  
In this paper we review the recent work on relational feature learning, and we provide an analysis of the role that multiplicative interactions play in learning to encode relations.  ...  Recently, there has been an increasing interest in learning to infer correspondences from data using relational, spatio-temporal, and bi-linear variants of deep learning methods.  ...  Acknowledgements: This work was supported in part by the German Federal Ministry of Education and Research (BMBF) in the project 01GQ0841 (BFNT Frankfurt).  ... 
doi:10.1109/tpami.2013.53 pmid:23787339 fatcat:ehgjqhnlhnbv5ch7qm76uotvqy

Learning Deep Architectures for AI

Y. Bengio
2009 Foundations and Trends® in Machine Learning  
such as Restricted Boltzmann Machines, used to construct deeper models such as Deep Belief Networks.  ...  In addition to the difficulty of coming up with the appropriate intermediate abstractions, the number of visual and semantic categories (such as MAN) that we would like an "intelligent" machine to capture  ...  This research was performed thanks to the funding from NSERC, MITACS, and the Canada Research Chairs.  ... 
doi:10.1561/2200000006 fatcat:pqujlozkonasra65suwxpvvuou

Towards 3D object detection with bimodal deep Boltzmann machines over RGBD imagery

Wei Liu, Rongrong Ji, Shaozi Li
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
In this work, we propose a cross-modality deep learning framework based on deep Boltzmann Machines for 3D Scenes object detection.  ...  Nowadays, detecting objects in 3D scenes like point clouds has become an emerging challenge with various applications.  ...  Bimodal Feature Learning We adopt bimodal Boltzmann deep machines for bimodal feature learning.  ... 
doi:10.1109/cvpr.2015.7298920 dblp:conf/cvpr/LiuJL15 fatcat:hub4aowrmjguzapslc3qxlb6ea

DeepAMR for predicting co-occurrent resistance of Mycobacterium tuberculosis

2019 Bioinformatics  
The results showed that DeepAMR outperformed baseline model and four machine learning models with mean AUROC from 94.4% to 98.7% for predicting resistance to four first-line drugs [i.e. isoniazid (INH)  ...  We then proposed an end-to-end multi-task model with deep denoising auto-encoder (DeepAMR) for multiple drug classification and developed DeepAMR_cluster, a clustering variant based on DeepAMR, for learning  ...  Comparing with wrapper and filter methods for selecting features, a better ranking could be obtained by learning the importance of variables within the model.  ... 
doi:10.1093/bioinformatics/btz067 pmid:30689732 pmcid:PMC6748723 fatcat:2xcffz2eqjaxvbuqduzc6xl6yq

Attentional Neural Network: Feature Selection Using Cognitive Feedback [article]

Qian Wang, Jiaxing Zhang, Sen Song, Zheng Zhang
2014 arXiv   pre-print
Attentional Neural Network is a new framework that integrates top-down cognitive bias and bottom-up feature extraction in one coherent architecture.  ...  We obtain classification accuracy better than or competitive with state of art results on the MNIST variation dataset, and successfully disentangle overlaid digits with high success rates.  ...  During the top-down pass, b generates a gating vector g = σ(U · b) with the feedback weights U . g selects and de-selects the features by modifying hidden activation h g = h g, where means pair-wised multiplication  ... 
arXiv:1411.5140v1 fatcat:uszkmugcmzcyjplkxempqvcq74

A Review of Artificial Intelligence Algorithms Used for Smart Machine Tools

Chih-Wen Chang, Hau-Wei Lee, Chein-Hung Liu
2018 Inventions  
neural network (ANN) method for rolling element bearing fault diagnosis [69], the Pca-based feature selection scheme for machine defect classification [70], the support vector machine (SVM) approach to  ...  machinery with massive data [78], a rapid Fourier transform (STFT)-deep learning scheme for rolling bearing fault diagnosis [79], a new bearing condition recognition method based on multi-feature extraction  ...  For the learning statistical features, real-value Gaussian-Bernoulli restricted Boltzmann machines (GRBMs) were stacked to develop a Gaussian-Bernoulli deep Boltzmann machine. Later, Li et al.  ... 
doi:10.3390/inventions3030041 fatcat:6qrwhmrl2bfwrgmovqvsyx5p3y

Automatic Gating of Attributes in Deep Structure

Xiaoming Jin, Tao He, Cheng Wan, Lan Yi, Guiguang Ding, Dou Shen
2018 Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence  
This demands adaptively and jointly modeling of attributes and deep structure by carefully examining their relationship.  ...  In this paper, we move forward along this new direction by proposing a deep structure named Attribute Gated Deep Belief Network (AG-DBN) that includes a tunable attribute-layer gating mechanism and automatically  ...  Acknowledgments The work is supported by National Basic Research Program of China (2015CB352300), National Natural Science Foundation of China (61571269), and Key technology R&D Program of Liaoning and  ... 
doi:10.24963/ijcai.2018/319 dblp:conf/ijcai/JinHWYDS18 fatcat:o4gcv7l4wbby5m6x76ot2vqtha

Deep Learning of Representations: Looking Forward [article]

Yoshua Bengio
2013 arXiv   pre-print
Deep learning research aims at discovering learning algorithms that discover multiple levels of distributed representations, with higher levels representing more abstract concepts.  ...  or local minima, designing more efficient and powerful inference and sampling procedures, and learning to disentangle the factors of variation underlying the observed data.  ...  Acknowledgments The author is extremely grateful for the feedback and discussions he enjoyed with collaborators Ian Goodfellow, Guillaume Desjardins, Aaron Courville, Pascal Vincent, Roland Memisevic and  ... 
arXiv:1305.0445v2 fatcat:cyfgf5trljfopcjsapicf4ay3q
« Previous Showing results 1 — 15 out of 434 results