Filters








2,512 Hits in 4.3 sec

Deep ReLU Networks Have Surprisingly Few Activation Patterns [article]

Boris Hanin, David Rolnick
<span title="2019-10-20">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In ReLU networks, the number of activation patterns is one measure of expressivity; and the maximum number of patterns grows exponentially with the depth.  ...  In this paper, we show that the average number of activation patterns for ReLU networks at initialization is bounded by the total number of neurons raised to the input dimension.  ...  This reflects the fact that deep ReLU nets may have many more activation regions than shallow networks with the same number of neurons.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1906.00904v2">arXiv:1906.00904v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/u55hw4vr5fa3rm2jo5zyitkywa">fatcat:u55hw4vr5fa3rm2jo5zyitkywa</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200726134007/https://arxiv.org/pdf/1906.00904v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1906.00904v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Deep BLSTM neural networks for unconstrained continuous handwritten text recognition

Volkmar Frinken, Seiichi Uchida
<span title="">2015</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/bytixui4szbgzkqzgwf65jphty" style="color: black;">2015 13th International Conference on Document Analysis and Recognition (ICDAR)</a> </i> &nbsp;
The second one are deep learning techniques, which greatly increased the performance of neural networks, by making use of many hidden layers.  ...  Extensive experimental evaluation on the IAM database demonstrate an increase of the recognition performance when using deep learning approaches over commonly used BLSTM neural networks, as well as insight  ...  RELATED WORKS Neural network-based pattern recognition has been given a large amount of attention in the last few years.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/icdar.2015.7333894">doi:10.1109/icdar.2015.7333894</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/icdar/FrinkenU15.html">dblp:conf/icdar/FrinkenU15</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/jfij23l6wzgthfkkbjehek5yui">fatcat:jfij23l6wzgthfkkbjehek5yui</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20170808120341/http://human.ait.kyushu-u.ac.jp/~uchida/Papers/ICDAR2015-Frinken.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a8/7b/a87b43c547200460030cce155d06ea997e19b590.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/icdar.2015.7333894"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

Implicit Rugosity Regularization via Data Augmentation [article]

Daniel LeJeune, Randall Balestriero, Hamid Javadi, Richard G. Baraniuk
<span title="2019-10-10">2019</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Deep (neural) networks have been applied productively in a wide range of supervised and unsupervised learning tasks.  ...  Unlike classical machine learning algorithms, deep networks typically operate in the overparameterized regime, where the number of parameters is larger than the number of training data points.  ...  Network with continuous piecewise affine activations We are primarily interested in deep networks constructed using piecewise affine activations (e.g., ReLU, leaky ReLU, absolute value).  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1905.11639v3">arXiv:1905.11639v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/fcer6nqw75cy5n2g7kijt5jema">fatcat:fcer6nqw75cy5n2g7kijt5jema</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200930231239/https://arxiv.org/pdf/1905.11639v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/a7/d1/a7d1fdc2b9d67e3bedcd9388862a129c8b9e67c5.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1905.11639v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Fast-classifying, high-accuracy spiking deep networks through weight and threshold balancing

Peter U. Diehl, Daniel Neil, Jonathan Binas, Matthew Cook, Shih-Chii Liu, Michael Pfeiffer
<span title="">2015</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/qm5nunzmyva4tfjekdcm34uvhq" style="color: black;">2015 International Joint Conference on Neural Networks (IJCNN)</a> </i> &nbsp;
To overcome the large computational cost of deep networks, spiking deep networks have recently been proposed, given the specialized hardware now available for spiking neural networks (SNNs).  ...  The presented analysis and optimization techniques boost the value of spiking deep networks as an attractive framework for neuromorphic computing platforms aimed at fast and efficient pattern recognition  ...  In recent years, spiking deep neural networks have become an increasingly active field of research.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/ijcnn.2015.7280696">doi:10.1109/ijcnn.2015.7280696</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/ijcnn/DiehlNB0LP15.html">dblp:conf/ijcnn/DiehlNB0LP15</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/qdomfyjbkvb6nkkytlzfb3thrq">fatcat:qdomfyjbkvb6nkkytlzfb3thrq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20170808105548/http://www.peterudiehl.com/Diehl_et_al_2015_IJCNN.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/8f/bf/8fbfb1e550c2bf2b65de54e2ae2ad05c9d9496fe.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/ijcnn.2015.7280696"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

Deep Learning Model for Sentiment Analysis on Short Informal Texts

Sam Farisa Chaerul Haviana, Bagus Satrio Waluyo Poetro
<span title="2022-02-16">2022</span> <i title="IAES Indonesia Section"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/myj7vzp6s5dvratkkqupv2ibgq" style="color: black;">Indonesian Journal of Electrical Engineering and Informatics (IJEEI)</a> </i> &nbsp;
This work focuses on building suitable preprocessing algorithm and building a simple deep learning network.  ...  The preprocessing algorithm was based on multiple word n-gram and Term Frequency-Inverse Document Frequency (TF-IDF) vectorization, and the network was built with a relatively shallow network.  ...  The network built in this work consists of Dense / Fully Connected layers, Leaky ReLU Activation [15] , Dropout layers, and a PReLU Activation [16] .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.52549/ijeei.v10i1.3181">doi:10.52549/ijeei.v10i1.3181</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/p4znpeekcrb2pjfxu2b75cbjhu">fatcat:p4znpeekcrb2pjfxu2b75cbjhu</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20220521075935/http://section.iaesonline.com/index.php/IJEEI/article/download/3181/674" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/50/95/5095ad3197212c26306e59d3967957defc5c6363.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.52549/ijeei.v10i1.3181"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> Publisher / doi.org </button> </a>

DNNs as Layers of Cooperating Classifiers

Marelie Davel, Marthinus Theunissen, Arnold Pretorius, Etienne Barnard
<span title="2020-04-03">2020</span> <i title="Association for the Advancement of Artificial Intelligence (AAAI)"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/wtjcymhabjantmdtuptkk62mlq" style="color: black;">PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE</a> </i> &nbsp;
We demonstrate intriguing regularities in the activation patterns of the hidden nodes within fully-connected feedforward networks.  ...  By tracing the origin of these patterns, we show how such networks can be viewed as the combination of two information processing systems: one continuous and one discrete.  ...  In this work we presented interesting regularities in the class-related activation patterns of nodes within a deep ReLU-activated network.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1609/aaai.v34i04.5782">doi:10.1609/aaai.v34i04.5782</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/khiq43fshzaivd4gb6b7ef4a5y">fatcat:khiq43fshzaivd4gb6b7ef4a5y</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20210716142148/https://repository.nwu.ac.za/bitstream/handle/10394/33960/davel-2020-dnns-layers-classifiers.pdf;jsessionid=A224356EAB418938231ACCCF1E4ECCE5?sequence=5" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/9f/64/9f647799baf415a7e6f81c440522a146df7dfad4.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1609/aaai.v34i04.5782"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> Publisher / doi.org </button> </a>

DNNs as Layers of Cooperating Classifiers [article]

Marelie H. Davel, Marthinus W. Theunissen, Arnold M. Pretorius, Etienne Barnard
<span title="2020-01-17">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
We demonstrate intriguing regularities in the activation patterns of the hidden nodes within fully-connected feedforward networks.  ...  A robust theoretical framework that can describe and predict the generalization ability of deep neural networks (DNNs) in general circumstances remains elusive.  ...  Figure 1 shows an example of the activation patterns that we have observed in numerous ReLU-activated networks of various architectures, trained with different algorithms on different classification tasks  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2001.06178v1">arXiv:2001.06178v1</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/qbmzgya6vrbhfcsoxr4qgn77he">fatcat:qbmzgya6vrbhfcsoxr4qgn77he</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200321102907/https://arxiv.org/pdf/2001.06178v1.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/99/f2/99f2801ed2bbc11d27ac1f424e6bcf79775dadae.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/2001.06178v1" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Deep Roots: Improving CNN Efficiency with Hierarchical Filter Groups

Yani Ioannou, Duncan Robertson, Roberto Cipolla, Antonio Criminisi
<span title="">2017</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/ilwxppn4d5hizekyd3ndvy2mii" style="color: black;">2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)</a> </i> &nbsp;
We propose a new method for creating computationally efficient and compact convolutional neural networks (CNNs) using a novel sparse connection structure that resembles a tree root.  ...  This allows a significant reduction in computational cost and number of parameters compared to state-of-the-art deep CNNs, without compromising accuracy, by exploiting the sparsity of inter-layer filter  ...  The motivation is that most of the convolutional filters respond to localized patterns in a small receptive field, with few requiring a larger receptive field.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvpr.2017.633">doi:10.1109/cvpr.2017.633</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/cvpr/IoannouRCC17.html">dblp:conf/cvpr/IoannouRCC17</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/arklgfsjuvd6jm2s7d6xjtupia">fatcat:arklgfsjuvd6jm2s7d6xjtupia</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20171209082432/http://openaccess.thecvf.com:80/content_cvpr_2017/papers/Ioannou_Deep_Roots_Improving_CVPR_2017_paper.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/22/5a/225a7dd3ebf4fe92cfdc388aff1a32f88497c8f5.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvpr.2017.633"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

Do Deep Neural Networks Learn Facial Action Units When Doing Expression Recognition? [article]

Pooya Khorrami, Tom Le Paine, Thomas S. Huang
<span title="2017-03-16">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Despite being the appearance-based classifier of choice in recent years, relatively few works have examined how much convolutional neural networks (CNNs) can improve performance on accepted expression  ...  We then qualitatively analyze the network by visualizing the spatial patterns that maximally excite different neurons in the convolutional layers and show how they resemble Facial Action Units (FAUs).  ...  A few works based on unsupervised deep learning have also tried to analyze the relationship between FAUs and the learned feature representations.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1510.02969v3">arXiv:1510.02969v3</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/woq3hfi765fadebaym5cah5sjy">fatcat:woq3hfi765fadebaym5cah5sjy</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200831084359/https://arxiv.org/pdf/1510.02969v3.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/b2/28/b228d1a12ea139cdc2c37170a014b4be9ff41bdd.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1510.02969v3" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Local Binary Convolutional Neural Networks

Felix Juefei-Xu, Vishnu Naresh Boddeti, Marios Savvides
<span title="">2017</span> <i title="IEEE"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/ilwxppn4d5hizekyd3ndvy2mii" style="color: black;">2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)</a> </i> &nbsp;
The design principles of LBC are motivated by local binary patterns (LBP).  ...  The linear weights combine the activated filter responses to approximate the corresponding activated filter responses of a standard convolutional layer.  ...  Deep convolutional neural networks (CNN), in particular, have enjoyed huge success in tackling many computer vision problems over the past few years, thanks to the tremendous development of many effective  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvpr.2017.456">doi:10.1109/cvpr.2017.456</a> <a target="_blank" rel="external noopener" href="https://dblp.org/rec/conf/cvpr/Juefei-XuBS17.html">dblp:conf/cvpr/Juefei-XuBS17</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/7v3ctws4bfbarb5ddztzn67qzq">fatcat:7v3ctws4bfbarb5ddztzn67qzq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20190819022256/http://openaccess.thecvf.com:80/content_cvpr_2017/papers/Juefei-Xu_Local_Binary_Convolutional_CVPR_2017_paper.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/dc/1d/dc1d544f73030726d77a5266176c686c44b0fa4b.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1109/cvpr.2017.456"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> ieee.com </button> </a>

Explaining the unique nature of individual gait patterns with deep learning

Fabian Horst, Sebastian Lapuschkin, Wojciech Samek, Klaus-Robert Müller, Wolfgang I. Schöllhorn
<span title="2019-02-20">2019</span> <i title="Springer Nature"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/tnqhc2x2aneavcd3gx5h7mswhm" style="color: black;">Scientific Reports</a> </i> &nbsp;
Machine learning (ML) techniques such as (deep) artificial neural networks (DNN) are solving very successfully a plethora of tasks and provide new predictive models for complex physical, chemical, biological  ...  The present paper studies the uniqueness of individual gait patterns in clinical biomechanics using DNNs.  ...  No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. This work was supported by the Brain Korea  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/s41598-019-38748-8">doi:10.1038/s41598-019-38748-8</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/30787319">pmid:30787319</a> <a target="_blank" rel="external noopener" href="https://pubmed.ncbi.nlm.nih.gov/PMC6382912/">pmcid:PMC6382912</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/vaw2lbhzmjagtlsc35263ju2eq">fatcat:vaw2lbhzmjagtlsc35263ju2eq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200826091613/https://arxiv.org/pdf/1808.04308v1.pdf" title="fulltext PDF download [not primary version]" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <span style="color: #f43e3e;">&#10033;</span> <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/8f/6f/8f6f7042b8897bacc287bb17e1977fd1b5c189c5.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/s41598-019-38748-8"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="unlock alternate icon" style="background-color: #fb971f;"></i> Publisher / doi.org </button> </a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6382912" title="pubmed link"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> pubmed.gov </button> </a>

Local Binary Convolutional Neural Networks [article]

Felix Juefei-Xu, Vishnu Naresh Boddeti, Marios Savvides
<span title="2017-07-01">2017</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
The design principles of LBC are motivated by local binary patterns (LBP).  ...  The linear weights combine the activated filter responses to approximate the corresponding activated filter responses of a standard convolutional layer.  ...  Deep convolutional neural networks (CNN), in particular, have enjoyed huge success in tackling many computer vision problems over the past few years, thanks to the tremendous development of many effective  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1608.06049v2">arXiv:1608.06049v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/zcojdqry3jexxbffepbcbzvoqq">fatcat:zcojdqry3jexxbffepbcbzvoqq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20191016014611/https://arxiv.org/pdf/1608.06049v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/18/b8/18b8871c2af635d1cdfe09642ec6581775cd3a10.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1608.06049v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Dissecting Deep Neural Networks [article]

Haakon Robinson, Adil Rasheed, Omer San
<span title="2020-01-19">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
In exchange for large quantities of data and processing power, deep neural networks have yielded models that provide state of the art predication capabilities in many fields.  ...  This work presents a novel algorithm that can compute the piecewise affine form of any fully connected neural network with rectified linear unit activations.  ...  The set of activations of all neurons in a network is called the activation pattern [25] , denoted by π.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1910.03879v2">arXiv:1910.03879v2</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/kzgs4za5hrcazptu6aufob4cxq">fatcat:kzgs4za5hrcazptu6aufob4cxq</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200929185352/https://arxiv.org/pdf/1910.03879v2.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/37/b5/37b5e3140f7b4c5ea9b9c5a3d42c19439177c8e3.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1910.03879v2" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>

Deep learning

Yann LeCun, Yoshua Bengio, Geoffrey Hinton
<span title="">2015</span> <i title="Springer Nature"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/drfdii35rzaibj3aml5uhvr5xm" style="color: black;">Nature</a> </i> &nbsp;
New learning algorithms and architectures that are currently being developed for deep neural networks will only accelerate this progress.  ...  Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.  ...  In the first layer, each word creates a different pattern of activations, or word vectors (Fig. 4) .  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/nature14539">doi:10.1038/nature14539</a> <a target="_blank" rel="external noopener" href="https://www.ncbi.nlm.nih.gov/pubmed/26017442">pmid:26017442</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/fi3tihaulbhtpjtapnixaz43zi">fatcat:fi3tihaulbhtpjtapnixaz43zi</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20180921223833/http://wvuieeepes.org:80/ee335/papers/nature14539.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/15/cf/15cf63f8d44179423b4100531db4bb84245aa6f1.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1038/nature14539"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> nature.com </button> </a>

Identity Crisis: Memorization and Generalization under Extreme Overparameterization [article]

Chiyuan Zhang and Samy Bengio and Moritz Hardt and Michael C. Mozer and Yoram Singer
<span title="2020-01-09">2020</span> <i > arXiv </i> &nbsp; <span class="release-stage" >pre-print</span>
Deeper CNNs often fail, but nonetheless do astonishing work to memorize the training output: because CNN biases are location invariant, the model must progressively grow an output pattern from the image  ...  The trained networks stereotypically take one of two forms: the constant function (memorization) and the identity function (generalization).  ...  H0-H9 show fully connected networks (ReLU activation) with the corresponding number of hidden layers.  ... 
<span class="external-identifiers"> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.04698v4">arXiv:1902.04698v4</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/y36l5agdcbbd3dx52dfvrpfmlm">fatcat:y36l5agdcbbd3dx52dfvrpfmlm</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20200321030423/https://arxiv.org/pdf/1902.04698v4.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] </button> </a> <a target="_blank" rel="external noopener" href="https://arxiv.org/abs/1902.04698v4" title="arxiv.org access"> <button class="ui compact blue labeled icon button serp-button"> <i class="file alternate outline icon"></i> arxiv.org </button> </a>
&laquo; Previous Showing results 1 &mdash; 15 out of 2,512 results