Intro Machine Learning

Fraida Fund

These materials are for your personal use. Please do not redistribute (e.g. post online) without permission.

Week 0: Course Intro

• Notes: FAQ for in person section
• Notes: FAQ for online section with scheduled meeting time
• Notes: FAQ for online section with no scheduled meeting time
• Notebook: Python + numpy tutorial
• Notebook: Colab tutorial
• Notes: Prerequisite review
• Notes: Textbooks and Resources

Week 1: Intro ML

• Notes: This Date in History - 1964 World’s Fair
• Handout: Intro to Machine Learning
• Handout: Explore your data
• Notebook: Exploratory data analysis

Week 2: Linear regression

• Notes: “Beauty in the Classroom” preview
• Handout: Linear regression
• Notes: Extended derivation of linear regression parameters
• Notebook: Compute regression coefficients by hand
• Notebook: Linear regression in depth

• Notebook: Case study on “Beauty in the Classroom”, Part 1

• Notebook: Linear regression on Advertising data (homework)

Week 3: Gradient descent, bias-variance tradeoff

• Notes: videos on convexity, directional derivatives, gradient descent
• Handout: Gradient descent

• Notebook: Gradient descent in depth

• Notes: Predicting the course of COVID-19 with a “cubic model”
• Handout: Error decomposition
• Notes: bias and variance for linear regression
• Notebook: Bias-variance tradeoff in depth

Week 4: Model selection, regularization

• Handout: Model selection
• Notebook: Model selection in depth

• Notebook: Case study on “Beauty in the Classroom”, Part 2

• Notebook: Model order selection for neural data (homework)

• Handout: Regularization
• Notebook: Regularization in depth

Week 5: Logistic regression for classification

• Notes: Classifier metrics
• Handout: Logistic regression
• Notebook: Logistic regression in depth
• Notebook: Logistic regression for handwritten digits classification
• Outside reference: Machine bias
• Outside reference: Can you make AI fairer than a judge?

• Notebook: COMPAS case study

• Notebook: Classify your own handwritten digit (homework)

Week 6: K nearest neighbor, feature selection

• Handout: K nearest neighbor
• Handout: Feature selection
• Notebook: K nearest neighbor in depth
• Notebook: Voter classification with K nearest neighbor (homework)

Week 7: Decision trees, ensembles

• Handout: Decision trees
• Handout: Ensemble methods
• Notebook: Decision trees and ensembles in depth
• Notebook: Adaboost in depth
• Notebook: Bias and variance of non-parametrics models

Week 8: Support vector classifiers, kernels

• Handout: Support vector classifier
• Handout: SVM with kernels
• Handout: Hyperparameter optimization
• Notebook: Support vector machines in depth
• Notebook: Bias and variance of SVM

• Notebook: Handwritten digits classification

• Notebook: UAV assisted wireless localization with Bayesian Optimization and Gaussian Process Regression (homework) (requires Hello, AERPAW)

Week 9: Neural networks

• Handout: Neural networks
• Notebook: Draw your own classification problem for a neural network
• Notebook: Draw your own classification problem for a neural network (Pytorch version)
• Notebook: Backpropagation
• Notebook: Neural networks for music classification (homework)

Week 10: Deep neural networks, convolutional neural networks

• Handout: Deep neural networks
• Handout: Convolutional neural networks
• Notebook: the Slash dataset
• Notebook: Convolutional neural networks
• Notebook: Transfer learning on rock, paper, scissors
• Notebook: Deep neural nets from 1989

Week 11: Deploying machine learning systems

• Outside reference: Using Deep Learning at Scale in Twitter’s Timelines
• Handout: Deploying machine learning systems
• Activity: Deploying machine learning systems on Kubernetes

Week 12: Unsupervised learning

• Handout: Unsupervised learning
• Notebook: Clustering
• Notebook: PCA on handwritten digits
• Notebook: PCA on rock, paper, scissors

Week 13: Reinforcement learning

• Handout: Reinforcement learning
• Notebook: Reinforcement learning