Deep Learning Prerequisites: Logistic Regression in Python
This course is a lead-in to deep learning and neural networks – it covers a popular and fundamental technique used in machine learning, data science and statistics: logistic regression. We cover the theory from the ground up: derivation of the solution, and applications to real-world problems. We show you how one might code their own logistic regression module in Python.
This course does not require any external materials. Everything needed (Python, and some Python libraries) can be obtained for free.
This course provides you with many practical examples so that you can really see how deep learning can be used on anything. Throughout the course, we’ll do a course project, which will show you how to predict user actions on a website given user data like whether or not that user is on a mobile device, the number of products they viewed, how long they stayed on your site, whether or not they are a returning visitor, and what time of day they visited.
Another project at the end of the course shows you how you can use deep learning for facial expression recognition. Imagine being able to predict someone’s emotions just based on a picture!
If you are a programmer and you want to enhance your coding abilities by learning about data science, then this course is for you. If you have a technical or mathematical background, and you want use your skills to make data-driven decisions and optimize your business using scientific principles, then this course is for you.
This course focuses on “how to build and understand“, not just “how to use”. Anyone can learn to use an API in 15 minutes after reading some documentation. It’s not about “remembering facts”, it’s about “seeing for yourself” via experimentation. It will teach you how to visualize what’s happening in the model internally. If you want more than just a superficial look at machine learning models, this course is for you.
HARD PREREQUISITES / KNOWLEDGE YOU ARE ASSUMED TO HAVE:
calculus
linear algebra
probability
Python coding: if/else, loops, lists, dicts, sets
Numpy coding: matrix and vector operations, loading a CSV file
TIPS (for getting through the course):
Watch it at 2x.
Take handwritten notes. This will drastically increase your ability to retain the information.
Write down the equations. If you don’t, I guarantee it will just look like gibberish.
Ask lots of questions on the discussion board. The more the better!
Realize that most exercises will take you days or weeks to complete.
Write code yourself, don’t just sit there and look at my code.
WHAT ORDER SHOULD I TAKE YOUR COURSES IN?:
Check out the lecture “What order should I take your courses in?” (available in the Appendix of any of my courses, including the free Numpy course)
Start Here
This lecture will outline what will be learned in the course. I explain the importance of knowing the math, and provide short descriptions of each section later covered. Feel free to skip it and jump right into the 3rd video.
Difference between supervised and unsupervised learning. Difference between classification and regression. Feel free to skip this one if you already know this.
An easy first quiz
Basics: What is linear classification? What's the relation to neural networks?
I discuss what linear classification is from a general standpoint, without invoking any specifics related to logistic regression. I provide a 2-dimensional binary classification example and go over how we would classify data into 1 of 3 cases: positive class, negative class, and don't know / not sure.
In this lecture I discuss a brief history of neural networks, and talk about how the characteristics of the neuron (action potential, signal propagation, inhibitory and excitatory behavior) are modeled in different ways: the Hodgkin Huxley mdoel, the FitzHugh Nagumo model, and the logistic model.
I show the feedforward calculation for the output of a logistic unit.
I show how to code the feedforward calculation for the output of a logistic unit in Python and numpy.
Solving for the optimal weights
I show how we can solve for the weights in a logistic regression model if we make assumptions about the distributions of the input data (equal variance, Gaussian-distributed).
All these symbols can get confusing so this is a lecture to give a short and simple description of what each letter "stands for". These are important to get right now, because we'll use it for this course and every later course on deep learning and machine learning.
I show the cross-entropy error formula, and describe why this is used as an appropriate objective function for logistic regression.
I show how to calculate the cross-entropy error of your model in Python and numpy.
I show how to derive the likelihood and log-likelihood of your model and data, and I show how maximizing the likelihood is equivalent to minimizing the cross-entropy.
I explain how gradient descent can be used to solve for the minima of a function, and the algorithm that can be used to iteratively update the logistic weights.
I show how to code weight updates for logistic regression using gradient descent in Python and numpy.
Practical concerns
I show how to apply regularization for logistic regression in Python.
I show how logistic regression can be used to solve the donut problem (where one class is circular and is inside another circular class).
This lecture describes how to apply logistic regression to the XOR problem by making it a 3-dimensional problem.
Checkpoint and applications: How to make sure you know your stuff
This is a clip from my natural language processing course. It shows you how to apply logistic regression to sentiment analysis - measuring how positive or negative a word is.
Project: Facial Expression Recognition
Appendix
