Lately, I’ve been using ChatGPT and Copilot for a variety of tasks ranging from understanding technical concepts to getting career advice. It’s been around two years since I started using ChatGPT, and while I’ve become adept at crafting detailed prompts to get the most out of these tools, I still have a beginner-level understanding of how large language models (LLMs) actually work.
This realization has led me to embark on a new journey: I aim to attain an intermediate to advanced understanding of machine learning technologies. My detailed plan starts with learning the foundational math of machine learning. Concepts such as single-variable calculus, linear algebra, probability and statistics, and a bit of multivariable calculus are super important to know. I’ve read that you can skip this part of the journey, but I’d rather not take any shortcuts and prefer to start from scratch. It’s been a while since I worked on any math-related topics, but I don’t mind revisiting them.
I plan on learning all these math topics from Math Academy, which is an amazing resource that I might write a separate blog post about. After the math stage, I plan on focusing on classical machine learning, where I’ll learn how to code basic ML algorithms from scratch. This includes basic regression and classification models, progressing up to small multi-layer neural networks. This stage is important because coding basic ML algorithms without any assistance can help develop the intuition needed to understand more advanced algorithms that will eventually come out.
Much of modern machine learning is built upon deep learning, so the next stage involves diving into deep learning, where I’ll learn about multi-layer neural networks with many parameters.
I’m trying to keep this blog short and concise, but those three stages will consist of a lot of reading, course material, programming projects, and tons of math!
Stage 1: Foundational Math
Go through the Foundation Series (I, II, III) courses and finish with the Mathematics for Machine Learning course using Math Academy
Stage 2: Classical Machine Learning
Learn how to code basic ML algorithms from scratch, including basic regression and classification models, progressing up to small multi-layer neural networks, using these resources Intro to Machine Learning & Algorithms and Stanford - Machine Learning (Andrew Ng)
- Course 1: Supervised Machine Learning: Regression and Classification
- Course 2: Advanced Learning Algorithms
- Course 3: Unsupervised Learning, Recommenders, Reinforcement Learning
Stage 3: Deep Learning
Learn about multi-layer neural networks with many parameters, using these resources
I’m excited to become an active member of the AI community that’s shaping the future. Whether it’s collaborating on models, datasets, and applications on platforms like Hugging Face or building models for personal projects, I’m eager to contribute. I’ll be sharing updates every month, so stay tuned!