Day 6: Logistic Regression Sample Problem, Foundations of Micrograd
Agenda
- 3:45-4:00pm: Debrief at tables
- 4:00-4:25pm: Logistic Regression Example Problem
- 4:25-5:05pm: Logistic Regression Learning Rule
- 5:05-5:25pm: Foundations of Micrograd
Debrief on the last assignment (15 minutes)
Warm up your brains by refreshing on the last assignment. As a reminder, some of the big ideas were classification algorithms, log loss, and confounding variables.
Logistic Regression Example Problem (25 minutes)
We’re now going to be diving into logistic regression. We’ll start out by writing the basic ideas of logistic regression up on the board, and we’ll go through a notebook that shows a sample problem.
The content we’re going to use for this is contained in the beginning part of assignment 6. Let’s jump over there and look at it together.
Next, we’re going to go through a Colab notebook that shows an example logistic regression problem.
Logistic Regression Learning Rule (40 minutes)
Let’s use assignment 6 to begin to unpack some of the concepts behind choosing the best set of weights for logistic regression. Before we start, we’ll go over our high-level strategy.
Foundations of Micrograd (rest of class)
Let’s use assignment 6 to build some of the foundations we’re going to need to optimize a wide range of machine learning models given a training set.