Deep learning: Technical introduction

September 05, 2017 · Declared Dead · 🏛 arXiv.org

Repo contents: .gitattributes, .gitignore, Acknowledgements.tex, AlexNet.pdf, Bottleneck.pdf, Bottleneck_BN.pdf, Bottleneck_BN_2.pdf, Bottleneck_BN_backprop.pdf, Bottleneck_BN_backprop_2.pdf, CNN_MM_pixels.pdf, CNN_MM_unpixels.pdf, Conclusion.tex, Conv_equiv.pdf, DEEP_LEARNING.bib, ELU.pdf, GoogleNet.pdf, Inception.pdf, Introduction.tex, LSTM_structure-peephole.pdf, LSTM_structure-tot.pdf, LSTM_structure.pdf, LeNet.pdf, Mediamobile.png, Preface.tex, README.md, RNN_structure-tot.pdf, RNN_structure.pdf, ReLU.pdf, ResNet.pdf, S_FNN.pdf, ThesisStyle.cls, VGG-conv.pdf, VGG-fc.pdf, VGG-pool-fc.pdf, VGG-pool.pdf, VGG.pdf, White_book-blx.bib, White_book.bcf, White_book.dvi, White_book.pdf, White_book.tex, chapter1.tex, chapter2.tex, chapter3.tex, conv_2d-crop.pdf, conv_4d-crop.pdf, cover_page-crop.pdf, fc_equiv.pdf, fc_resnet.pdf, fc_resnet_2.pdf, fc_resnet_3.pdf, formatAndDefs.tex, fully_connected.pdf, input_layer.pdf, lReLU.pdf, mathpazo.sty, output_layer.pdf, padding.pdf, pagedecouv.sty, pool_4d-crop.pdf, sigmoid.pdf, softmax.pdf, tanh.pdf, tanh2.pdf

Authors Thomas Epelbaum arXiv ID 1709.01412 Category stat.ML: Machine Learning (Stat) Cross-listed cs.LG Citations 26 Venue arXiv.org Repository https://github.com/tomepel/Technical_Book_DL ⭐ 1388 Last Checked 1 month ago

Abstract

This note presents in a technical though hopefully pedagogical way the three most common forms of neural network architectures: Feedforward, Convolutional and Recurrent. For each network, their fundamental building blocks are detailed. The forward pass and the update rules for the backpropagation algorithm are then derived in full.