In-Depth Machine Learning for Teens: Introduction
Welcome to my series!
Here, you can learn the mathematics and intuition behind machine learning, building from the bottom up — not only that but everything is mathematically simplified to a level anyone can understand.
Why I Made This Series
Ever since 7th grade, I’ve wanted to learn about machine learning and AI. At that point, I had both limited programming and mathematical skills and was unable to get very far. However, after a year of practice, I got to the point where I was able to write code pretty well. Then, I retried my hand at machine learning.
Unfortunately, I was unable to get very far, as the majority of resources that I found simply used a library that did everything for you in a very abstract fashion. Of course, learning how to use a library would have allowed me to apply machine learning for practical uses. However, as a naturally curious person, I wasn’t satisfied with just using something by assuming it works — I was interested in learning how machine learning worked on the inside.
This led me to read a bunch of articles and papers about how machine learning actually worked on the inside. At this point, I was a rising junior. Once again, I was unable to get very far, as I did not have the background knowledge required to understand the material, as all of the resources assumed an undergrad audience or higher. However, I was able to get an intuitive understanding of the algorithms described.
Two years later, near the end of 10th grade, I achieved both an intuitive and mathematical understanding of calculus, after which I took another look at machine learning. This time, I was able to understand most of the material to a relatively high degree.
I had to follow a long and convoluted path in order to learn machine learning up to a point I was satisfied with it, but you don’t have to! My goal in this article series is to provide you with a mathematical and intuitive introduction to machine learning, made to be accessible to everyone.
Prerequisites
- Basic understanding of general mathematics (probability, algebra, manipulating equations and numbers, etc.)
- Basic python coding skills (general operations, loops, functions, how to use an imported module/library)
- An enthusiastic learning attitude
- Patience! Especially when we get to matrices!
What You Can Expect
- Articles and hands-on labs
- Detailed (and sometimes long) explanations of concepts
- Involved mathematics with intuitive explanations
- Terminology and optimizations
- Building everything from scratch
- General frameworks and instructions in labs
- Steps and methods to calculate gradients (~fancy calculus~) in labs
- Learning how to use the Numpy library
What We Will Not Cover
- How to utilize GPUs or TPUs to speed up training (especially for neural networks)
- Advanced neural network uses
- How to use machine learning libraries such as PyTorch, TensorFlow, SkLearn, etc.
- How to take over the world with AI
Now that we have all of that covered, let’s jump into the articles!
Table of Contents
- Introduction
- Gradient Descent
- Linear Regression
- Training Faster and Better
- Logistic Regression
- Neural Networks
Disclaimer
Please note that the majority of the media used for this series has been generated by me. Any images that I do not own have a link to their source underneath them. You can view the images and the code used to generate them at https://github.com/Endothermic-Dragon/Polygence.