Skip to content

Coursera Deep Learning Specialization: Code Implementation, Lecture Notes and Corresponding Papers

Notifications You must be signed in to change notification settings

MichaelYxWang/Coursera-Deep-Learning

Repository files navigation

Coursera-Deep-Learning

The goal of setting up this repo is to make full use of Dr Andrew Ng's Deep Learning Specialization.

This repo mainly provides the following features:

  1. For review purpose : A more convenient visualization of jupyter notebooks without setting up notebook server locally ✅
  2. The references of papers which appear in the 5 courses as well as some notes about the papers ✅
  3. Nicely commented code from helper functions to project architecture as well as a guideline of how to go through them (Ongoing👍)
  4. Extend the project to end to end system: from data labeling to research diary (Ongoing👍)
  5. Utilizae git lfs to store large resnet.h5 and vgg.mat file. Now every project should run in local without problems ✅
  6. (NEW from 2020!)To review the lecture notes more effectively, I organize all the concepts in the format of questions in order to build better deep learning fundations. If you want to collaborate on this, please message me :) (Ongoing👍)

Recourse collection contributors: Michael Wang, Richard Xu, Constantine Cheng and Jay Xiao

We have also included the brilliant Chinese notes written by Dr. Haiguang Huang to faciliate understanding of the material.

Course 4 Week 2 :

Case study - Classic Networks:

[1] "Gradient-based learning applied to document recognition"[pdf]

[2] "Very deep convolutional networks for large-scale image recognition"[pdf]

[3] "ImageNet classification with deep convolutional neural networks"[pdf]

ResNets

[4] "Deep residual networks for image recognition"[pdf]

Networks in networks

[5] "Network in network"[pdf]

Inception Network

[6] "Going deeper with convolutions"[pdf]

Course 4 Week 3 :

Convolutional Implementation of Sliding Windows

[7] "OverFeat: Integrated recognition, localization and detection using convolutional networks"[pdf]

Bounding Box Predictions, Anchor Boxes, YOLO Algorithm

[8] "You Only Look Once"[pdf]

Region Proposals

[9] "Rich feature hierarchies for accurate object detection and semantic segmentation"[pdf]

[10] "Fast R-CNN"[pdf]

[11] "Faster R-CNN: Towards real-time object detection with region proposal networks"[pdf]

Course 4 Week 4 :

Siamese network , Face Verification and Binary Classfication

[12] "DeepFace closing the gap to human level performance"[pdf]

Triplet Loss

[13] "FaceNet: A unified embedding for face recognition and clustering"[pdf]

What are deep ConvNets Learning?

[14] "Visualizing and understanding convolutional networks"[pdf]

Cost Function, Content Cost Function, Style Cost Function

[15] "A neural algorithm of artistic style"[pdf]

Course 5 Week 1 :

GRU (Gated Recurrent Unit)

[16] "On the Properties of Neural Machine Translation- Encoder–Decoder Approaches"[pdf]

[17] "Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling"[pdf]

LSTM (Long Short Term Memory)

[18] "Long Short Term Memory"[pdf]

Course 5 Week 2 :

Visualize word embeddings

[19] "Visualizing Data using t-SNE"[pdf]

Word embedding's relation to face encoding

[20] "DeepFace: Closing the Gap to Human-Level Performance in Face Verification"[pdf]

Properties of word embeddings

[21] "Linguistic Regularities in Continuous Space Word Representations"[pdf]

Neural Language Model

[22] "Linguistic Regularities in Continuous Space Word Representations"[pdf]

Word2Vec

[23] "Efficient Estimation of Word Representations in Vector Space"[pdf]

Negative sampling

[24] "Distributed Representations of Words and Phrases and their Compositionality"[pdf]

Glove word embedding

[25] "GloVe: Global Vectors for Word Representation"[pdf]

Debiasing word embeddings

[26] "Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings"[pdf]

Course 5 Week 3 :

Sequence to sequence models

[27] "Sequence to Sequence Learning with Neural Networks"[pdf]

[28] "Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation"[pdf]

Image Captioning

[29] "Deep Captioning with Multimodal Recurrent Neural Networks (m-RNN)"[pdf]

[30] "Show and Tell: A Neural Image Caption Generator"[pdf]

[31] "Deep Visual-Semantic Alignments for Generating Image Descriptions"[pdf]

Evaluate machine translation

[32] "BLEU: a Method for Automatic Evaluation of Machine Translation"[pdf]

Attention Model

[33] "Neural Machine Translation by Jointly Learning to Align and Translate"[pdf]

[34] "Show, Attend and Tell: Neural Image Caption Generation with Visual Attention"[pdf]

[35] "Attention Is All You Need"[pdf]

Speech Recognition and Trigger Word Detection

[36] "Connectionist Temporal Classification: Labelling Unsegmented Sequence Data with Recurrent Neural Networks"[pdf]