Course lists

Deep Neural Networks (DNNs) by Andrew Ng

  • 1 What is Deep Learning?
  • 2 What is a Neural Network?
  • 3 Supervised Learning with Neural Networks
  • 4 Drivers Behind the Rise of Deep Learning
  • 5 Binary Classification in Deep Learning
  • 6 Logistic Regression
  • 7 Logistic Regression Cost Function
  • 8 Gradient Descent
  • 9 Derivatives
  • 10 Derivatives Examples
  • 11 Computation Graph
  • 12 Derivatives with a Computation Graph
  • 13 Logistic Regression Derivatives
  • 14 Gradient Descent on m Training Examples
  • 15 Vectorization
  • 16 More Vectorization Examples
  • 17 Vectorizing Logistic Regression
  • 18 Vectorizing Logistic Regression’s Gradient Computation
  • 19 Broadcasting in Python
  • 20 Python-Numpy
  • 21 Jupyter-iPython
  • 22 Logistic Regression Cost Function Explanation
  • 23 Neural Network Overview
  • 24 Neural Network Representation
  • 25 Computing a Neural Network’s Output
  • 26 Vectorizing Across Multiple Training Examples
  • 27 Vectorized Implementation Explanation
  • 28 Activation Functions
  • 29 Why Non-Linear Activation Function?
  • 30 Derivatives of Activation Functions
  • 31 Gradient Descent for Neural Networks
  • 32 BackPropagation Intuition (Optional)
  • 33 Random Initialization of Weights
  • 34 Deep L-layer Neural Network
  • 35 Forward Propagation in Deep Networks
  • 36 Getting your Matrix Dimension Right
  • 37 Why DEEP representation?
  • 38 Building Blocks of Deep Neural Network
  • 39 Forward Propagation for Layer L
  • 40 Parameters vs Hyperparameters
  • 41 Brain and Deep Learning
  • 42 Train/Dev/Test sets
  • 43 Bias/ Variance
  • 44 Basic “Recipe” of Machine Learning
  • 45 Regularization
  • 46 Why Regularization reduces Overfitting?
  • 47 Dropout Regularization
  • 48 Why does drop-out work?
  • 49 Other Regularization Methods
  • 50 Normalizing Input
  • 51 Vanishing / Exploding Gradients
  • 52 Weight Initialization for deep networks
  • 53 Numerical Approximation of Gradients
  • 54 Gradient Checking
  • 55 Gradient Checking Implantation Notes
  • 56 Mini Batch Gradient Descent
  • 57 Understanding Mini-Batch Gradient Descent
  • 58 Exponentially Weighted Averages
  • 59 Understanding Exponentially Weighted Averages
  • 60 Bias Correction in Exponentially Weighted Average
  • 61 Gradient Descent with Momentum
  • 62 RMSprop
  • 63 Adam Optimization Algorithm
  • 64 Learning Rate Decay
  • 65 The Problem of Local Optima
  • 66 Tunning Process
  • 67 Right Scale for Hyperparameters
  • 68 Hyperparameters tuning in Practice: Panda vs Caviar
  • 69 Batch Norm
  • 70 Fitting Batch Norm into a Neural Network
  • 71 Why Does Batch Nom Work?
  • 72 Batch Norm at Test Time
  • 73 Softmax Regression
  • 74 Training a Softmax Classifier
  • 75 Deep Learning Frameworks
  • 76 TensorFlow
  • 77 Why ML Strategy?
  • 78 Orthogonalization
  • 79 Single Number Evaluation Metric
  • 80 Satisfying and Optimizing Metrics
  • 81 train/dev/test distributions
  • 82 Size of dev and test sets
  • 83 When to change dev/test sets and metrics?
  • 84 Why human-level performance?
  • 85 Avoidable Bias
  • 86 Understanding Human-Level Performance
  • 87 Surpassing Human-Level Performance
  • 88 Improving Your Model Performance
  • 89 Carrying Out Error Analysis
  • 90 Cleaning Up Incorrect Labeled Data
  • 91 Build Your First System Quickly, Then Iterate
  • 92 Training and Testing on Different Distributions
  • 93 Bias and Variance with Mismatched data distributions
  • 94 Addressing Data Mismatch
  • 95 Transfer Learning
  • 96 Multi-Task Learning
  • 97 End-to-End Deep Learning
  • 98 Whether to use End-to-End Learning

Convolutional Neural Networks (CNNs) by Andrew Ng

  • CNN1: What is Computer Vision?
  • CNN2: Edge Deletion Example
  • CNN3: More Edge Deletion
  • CNN4: Padding
  • CNN5: Strided Convolution
  • CNN6: Convolutions Over Volume
  • CNN7: One Layer of A Convolutional Network
  • CNN8: A Simple Convolution Network Example
  • CNN9: Pooling Layers
  • CNN10: Convolutional Neural Network Example
  • CNN11: Why Convolutions?
  • CNN12: Look at Case Studies
  • CNN13: Classic Networks
  • CNN14: Residual Networks (ResNets)
  • CNN15: Why Residual Network Works Well?
  • CNN16: Network in Network and 1*1 Convolutions
  • CNN17: Inception Network Motivation
  • CNN18: Inception Network
  • CNN19: ConvNets: Using open-Source Implementation
  • CNN20: Transfer Learning
  • CNN21: Data Augmentation
  • CNN22: The State of Computer Vision
  • CNN23: Object Detection: Object Localization
  • CNN24: Object Detection: Landmark Detection
  • CNN25: Object Detection: Object Detection
  • CNN26: Object Detection: Convolutional Implementation of Sliding Windows
  • CNN27: Object Detection: Bounding Box Predictions
  • CNN28: Object Detection: Intersection Over Union
  • CNN29: Object Detection: Non-max Suppression
  • CNN30: Object Detection: Anchor Boxes
  • CNN31: Object Detection: YOLO Algorithm
  • CNN32: Object Detection : Region Proposal (optional)
  • CNN33: Face Recognition: What is Face Recognition?
  • CNN34: Face Recognition, One-Shot Learning
  • CNN35: Face Recognition: Siamese Network
  • CNN36: Face Recognition: Triplet Loss
  • CNN37: Face Recognition: Face Verification and Binary Classification
  • CNN38: Neural Style Transfer: What is it?
  • CNN39: Neural Style Transfer: What are deep ConvNets Learning?
  • CNN40: Neural Style Transfer: Cost Function
  • CNN41: Neural Style Transfer : Content Cost Function
  • CNN42: Neural Style Transfer: Style Cost Function
  • CNN43: 1D and 3D Generalization of Models

Recurrent Neural Networks (RNNs) by Andrew Ng

  • RNN1: Why sequence models?
  • RNN2: Notation
  • RNN3: Recurrent Neural Network Model
  • RNN4: Backpropagation through time
  • RNN5: Different types of RNNs
  • RNN6: Language model and sequence generation
  • RNN7: Sampling novel sequences
  • RNN8: Vanishing gradients with RNNs
  • RNN9: Gated Recurrent Unit GRU
  • RNN10: LSTM long short term memory units
  • RNN11: Bidirectional RNN
  • RNN12: Deep RNNs
  • RNN13: Word representation
  • RNN14: Using word embeddings
  • RNN15: NLP Properties of word embeddings
  • RNN16: NLP Embedding matrix
  • RNN17: NLP Learning word embeddings
  • RNN18: NLP Word2Vec
  • RNN19: NLP Negetive sampling
  • RNN20: NLP - GloVe word vectors
  • RNN21: NLP - Sentiment classification
  • RNN22: NLP - Debiasing word embeddings
  • RNN23: Sequence to sequence models - Basic models
  • RNN24: Piccking the most likely sentence
  • RNN25: Beam search
  • RNN26: Refinement to beam search
  • RNN27: Error analysis on beam search
  • RNN28: Bleu score (optional)
  • RNN29: Attention model - intuition
  • RNN30: Attention model
  • RNN31: Audio data - Speech recognition
  • RNN32: Audio data -Trigger word detection
  • RNN33: Summary and Thank you!