MACHINE LEARNING SPECIALIZATION

• By Andrew Ng Offered by DeepLearning.AI

---

COURSE 1 : SUPERVISED ML REGRESSION & CLASSIFICATION

• Week 01 : Intro to ML
  1. Overview of ML
     • Learning Objective
     • Applications of ML
  2. Supervised vs Unsupervised ML
     • What is ML
     • Supervised Learning
     • Supervised Learning Types
     • Unsupervised Learning
  3. Regression Model
     • Linear Regression with one variable
     • Lab : Model Representation
     • Cost Function
     • Cost Function Intuition
     • Visualizing the cost function
     • Lab : Cost Function
  4. Training the model with GD
     • Gradient Descent
     • Implementing Gradient Descent
     • Gradient Descent Intuition
     • Learning Rate
     • Gradient Descent for Linear Regression
     • Running Gradient Descent
     • Lab : To automate the process of optimizing w and b using gradient descent
  5. Linear Regression from Scratch

• Week 02 : Regression with Multiple input variables
  1. Multiple Linear Regression
     • Multiple Features
     • Vectorization
     • Lab : Python, Numpy and Vectorization
     • Gradient Descent for Multiple Linear Regression
     • Lab : Multiple Linear Regression
     • Gradient Descent with Multiple Variables
  2. Gradient Descent in Practice
     • Feature Scaling
     • Mean Normalization
     • Z-score Normalization
     • Checking Gradient descent for convergence
     • Choosing the Learning Rate
     • Lab : Feature Scaling and Learning Rate (Multi-Variable)
     • Polynomial Regression
     • Lab : Feature Engineering & Polynomial Regression
     • Selecting Features
     • Scaling Features
     • Linear Regression using scikit-learn
  3. Practice Lab Linear Regression

• Week 03 : Classification
  1. Classification with Logistic Regression
     • Classification
     • Lab : Classification
     • Linear Regression Approach
     • Logisitic Regression
     • Lab : Logisitic Regression
     • Sigmoid or Logistic Function
     • Decision Boundary
     • Lab : Logistic Regression, Decision Boundary
  2. Cost Function for Logistic Regression
     • Logistic Loss Function
     • Lab : Logistic Regression, Logistic Loss
     • Simplified Cost Function for Logistic Regression
     • Lab : Cost Function for Logistic Regression
  3. Gradient Descent for Logistic Regression
     • Lab : Gradient Descent for Logistic Regression
     • Logistic Gradient Descent
     • Lab : Logistic Regression Using scikit-Learn
  4. The Problem of Overfitting
     • The Problem of Overfitting
     • Regularization to Reduce Overfitting
     • Lab : Overfitting
     • Cost Function with Regularization
     • Regularized Linear Regression
     • Regularizeed Logistic Regression
     • Lab : Regularized Cost and Gradient

Certificate of Completion

COURSE 2 : ADVANCED LEARNING ALGORITHMS

• Week 01 : Neural Network
  1. Neural Network Intuition
     • Neural Networks
     • Demand Prediction
     • Example : Recognizing images
  2. Neural Network Model
     • Neural Network Layer
     • More Complex Neural Networks
     • Forward Propagation
     • Handwritten digit recognition
     • Lab : Neurons and Layers
  3. Tensorflow Implementation
     • Inference in Code
     • Build the model using TensorFlow
     • Model for digit Classification
     • Data in Tensorflow
     • Feature vectors
     • Activation vector
     • Building a neural network architecture
     • Lab : Coffee Roasting in Tensorflow
  4. Neural Network Implmentation in Python
     • Forward Propagation in a single layer
     • Lab : Coffee Roasting Numpy
  5. Speculations on Artificial General Intelligence (AGI)
     • Artificial Narrow Intelligence (ANI)
     • Artificial General Intelligence (AGI)
  6. Vectorization
     • How neural networks are implemented efficiently.
     • Matrix Multiplication

• Week 02 : Neural Network Training
  1. Neural Network Training
     • Tensorflow implementation
     • Training Details
  2. Activation Function
     • Alternatives to the Sigmoid activation
     • Choosing Activation Functions
     • Why do we need activation functions
     • Lab : ReLu activation
  3. Multiclass Classification
     • Multiclass
     • softmax
     • Neural network with softmax output
     • Improved implementation of softmax
     • Classification with multiple outputs.
     • Lab : Softmax Function
     • Lab : Multiclass
  4. Additional Neural Network Concepts
     • Additional Optimization
     • Additional Layer Types
     • Convolutional layer
     • Convolutional Neural Network
  5. Back Propagation
     • What is a derivative
     • Computation Graph
     • Larger Neural Network

• Week 03 : Advice for Applying ML
  1. Advice for Applying Machine Learning
     • Deciding what to try next
     • Evaluating a model
     • Model Selection and Training/Cross validation/test sets
     • Lab : Model Evaluation and Selection
  2. Bias and Variance
     • Diagnosing bias and variance
     • Regularization and Bias/Variance
     • Establising a baseline level of performance
     • Learning Curves
     • Deciding what to try next revisited
     • Bias/Variance and Neural Networks
     • Lab : Diagnosing Bias and Variance
  3. Machine Learning Development Process
     • Iterative Loop of ML development
     • Error Analysis
     • Adding Data
     • Transfer Learning
     • Full cycle of a machine learning project
  4. Skewed Database
     • Error metrics for skewed datasets
     • Trading off precision and recall

• Week 04 : Decision Trees
  1. Decision Trees
     • Decision Trees
     • Learning Process
  2. Decision Tree Learning
     • Entropy as a measure of impurity
     • Choosing a split : Information Gain
     • Decision Tree Learning
     • Using one-hot encoding of categorial features
     • Continuous Valued Features
     • Regression Trees
     • Lab : Decision Trees
  3. Tree Ensembles
     • Using Multiple Decision Trees
     • Tree ensemble
     • Sampling with replacement
     • Random Forest Algorithm
     • XGBoost
     • When to use decision trees
     • Decision Trees vs Neural Networks
     • Lab : Tree Ensembles - Heart Failure Prediction

Certificate of Completion

COURSE 3 : UNSUPERVISED LEARNING, RECOMMENDERS, REINFORCEMENT LEARNING

• Week 01 : Unsupervised Learning
  1. Welcome
  2. Clustering
     • What is clustering
     • Applications of Clustering
     • K-means intuition
     • K-means Algorithm
     • Optimization Objective
     • Initializing K-means
     • Choosing the number of clusters
  3. Anomaly Detection
     • Finding unusual events
     • Anomaly Detection
     • Density estimation
     • Gaussian (normal) Detection
     • Anomaly Detection Algorithm
     • Developing and evaluating an anomaly detection system
     • Anomaly Detection vs Supervised Learning
     • Choosing what features to use
  4. Practice Lab : K-means Clustering
     • Image Compression
  5. Practice Lab : Anomaly Detection in server computers

• Week 02 : Recommender Systems
  1. Collaborative Filtering
     • Making Recommendations
     • Using per-item features
     • Collaborative Filtering Algorithm
  2. Recommender Systems Implementation
     • Mean Normalization
     • Tensorflow implementation of collaborative filtering
     • Finding Related items
  3. Content Based Filtering
     • Collaborative Filtering vs Content-base Filtering
     • Deeplearning for content-base filtering
     • Recommending from a large catalogue
     • Tensorflow implementation of content-based filtering
  4. Principal Component Analysis
     • Reducing the number of features
     • PCA algorithm
     • PCA in code
     • Applications of PCA
     • Lab : PCA and data visulization
  5. Practice Lab : Collaborative Filtering Recommender Systems
     • Recommender system for movies
  6. Practice Lab : Deep Learning for Content-based Filtering
     • Recommebder for movies

• Week 03 : Reinforcement Learning
  1. Reinforcement Learning Introduction
     • What is Reinforcement Learning
     • The return in Reinforcement Learning
     • Making decisions : Polices of reinforcement learning
     • The goal of reinforcement learning
     • Markov Decision Process(MDP)
  2. State-action value function
     • State-action value function definition (Q function)
     • Lab : State action value function
       • mars rover example
     • Bellman Equation
     • Random (stochastic) environment
  3. Continuous state space
     • Examples of continous state space applications
     • Lunar lander
     • Learning the state-function
     • Algorithm refinement : Improved neural netwprk architecture
     • Algorithm refinment e-greedy policy
     • Mini-batching and soft-updates
     • The state of reinforcement learning

Certificate of Completion