CS545: Machine Learning

This is an old revision of the document!

August

Week	Topic	Lecture Notes	Reading	Assignments
Week 1: Aug 20, 22	Course overview. Machine Learning and AI: History and Present Boom Jupyter notebooks.	01 Introduction to CS545 01a Simple Animations 02 Searching for Good Weights in a Linear Model	JupyterLab Introduction, watch the video then play with jupyter lab. What is Data Analysis? How to Visualize Data with Python, Numpy, Pandas, Matplotlib & Seaborn Tutorial, by Aakash NS	Not graded: Please fill out this anonymous survey before Thursday class.
Week 2: Aug 27, 29	Optimization algorithms. Simple linear and nonlinear models. Confidence intervals.	02 Searching for Good Weights in a Linear Model 02a Input Importance and Generative AI---Friend or Foe 03 Fitting Simple Models Using Gradient Descent in the Squared Error 04 Training Multiple Models to Obtain Confidence Intervals

September

Week	Topic	Lecture Notes	Reading	Assignments
Week 3: Sept 3, 5	Introduction to neural networks.	05 Introduction to Neural Networks	3Blue1Brown Introduction to Neural Networks in the first five chapters provides a fun video tutorial including error backpropagation.
Week 4: Sept 10, 12	Design of NeuralNetwork class. Optimizers. Overview of A2. Memory organization for neural network parameters. Optimizers tailored for neural networks.	06 Python Classes 07 Optimizers Simple 08 Collecting All Weights into One-Dimensional Vector for Use in Optimizers 08a Optimizers	Weight Initialization for Deep Learning Neural Networks, by Jason Brownlee	A1 due Monday, September 9th, 10:00 PM.
Week 5: Sept 17, 19 Chuck's office hours cancelled today.	Introduction to Classification.	09 Introduction to Classification		A2 NeuralNetwork Class due Thursday, September 19, 10:00 PM. Here is an example solution to A2: A2 NeuralNetwork Class Solution. Examples of good A2 solutions from your classmates can be found here
Week 6: Sept 24, 26	Classification with Logistic Regression.	10 Classification with Linear Logistic Regression 11 Classification with Nonlinear Logistic Regression Using Neural Networks

October

Week	Topic	Lecture Notes	Reading	Assignments
Week 7: Oct 1, 3	Classification with Nonlinear Logistic Regression, K-Nearest-Neighbors. Clustering with K-Means.	12 K-Means Clustering, K-Nearest-Neighbor Classification		A3 NeuralNetwork Class Using Optimizers due Friday, October 4th, 10:00 PM. A3grader.zip is now available. Examples of good A3 solutions from your classmates can be found here
Week 8: Oct 8, 10	A4. Introduction to Reinforcement Learning.	13 Introduction to Reinforcement Learning 14 Reinforcement Learning with Neural Networks as Q Function	John Hopfield and Geoffrey Hinton awarded Nobel Physics Prize Will digital intelligence replace biological intelligence? Geoffrey Hinton Warns of the “Existential Threat” of AI
Week 9: Oct 15, 17	Reinforcement learning with Q Function as Neural Network. Learning to play games.	14 Reinforcement Learning with Neural Networks as Q Function 15 Reinforcement Learning for Two Player Games 16 Targets and Deltas Summary		A4 Neural Network Classifier due Friday, October 18th, 10:00 PM. Examples of good A4 solutions from your classmates can be found here
Week 10: Oct 22, 24	Modular framework for reinforcement learning. Parallel processing with ray. Introductions to Pytorch and Jax.	17 Modular Framework for Reinforcement Learning 18 Ray for Parallel Processing 19 More Tic-Tac-Toe and a Simple Robot Arm 19 Introduction to Jax 21 Introduction to Pytorch	President Biden's Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence	Project proposal due at 10 pm Friday evening, October 25th.
Week 11: Oct 29, 31	A5. Pytorch. Convolutions.	22 Introduction to Convolutional Neural Networks 23 Convolutional Neural Network in Numpy	Backpropagation for Convolution by Mayank Kaushik Convolutions and Backpropagations by Pavithra Solai

November

Week	Topic	Lecture Notes	Reading	Assignments
Week 12: Nov 5, 7	Convolutional Neural Networks in Pytorch	24 Convolutional Neural Network Class in Pytorch 25 CNN on One-Dimensional Data		A5 Pole Balancing with Reinforcement Learning Due Friday, November 8th, 10:00 PM. Examples of good A5 solutions can be found here
Week 13: Nov 12, 14	Ensembles. Autoencoders. Recurrent Neural Networks.	26 Ensembles 27 Autoencoders 28 Recurrent Networks in Numpy 29 Recurrent Networks in Pytorch
Week 14: Nov 19, 21	Word Embeddings. Transformers. Support Vector Machines.	30 Word Embeddings 31 Introdcution to Transformers 32 Support Vector Machines	ChatGPT generates fake data set to support scientific hypothesis ClinicalBench: Can LLMs Beat Traditional ML Models in Clinical Prediction? F1 Score vs. Accuracy: Which Should You Use?	A6 Convolutional Neural Networks Due Thursday, November 21st, 10:00 PM.
Fall Break: Nov 25-29	No classes.

December

Week	Topic	Lecture Notes	Reading	Assignments
Week 15: Dec 3, 5	Transformers. Streamlit.
Dec 10-12	Final Exam Week	No Exams in this course

Table of Contents

August

September

October

November

December