I'm really interested in diving deep into AI and want to learn as if it's my major. I have a solid background in applied math and computer science from my undergraduate studies. The issue I'm facing is that most resources out there, especially on YouTube, seem to focus on quick results, like becoming an AI engineer in six months or creating LLMs, which isn't what I need. I'm looking for serious resources that cover the theory and fundamentals of AI rather than those quick, non-theoretical bootcamp courses.
6 Answers
With your strong math and computer science background, you're well-positioned to start studying AI rigorously. I recommend beginning with 'Artificial Intelligence: A Modern Approach' by Russell and Norvig. It’s the go-to textbook that covers everything from search algorithms to neural networks with a good depth of theory. For a focus on machine learning, check out 'Pattern Recognition and Machine Learning' by Bishop or 'Elements of Statistical Learning' by Hastie et al. Both dive deep into the math, which you'll appreciate.
You might also want to look into Stanford’s CS229 course, which you can access for free online. Andrew Ng does an excellent job of teaching the essentials, incorporating linear algebra and calculus. Additionally, MIT offers open courseware for their AI classes. I suggest following university syllabi from top schools like MIT, Stanford, or Berkeley to structure your learning.
For deep learning, 'Deep Learning' by Goodfellow is a must-read. As you get comfortable, pick a specific area that fascinates you and start delving into research papers on ArXiv. Try to implement algorithms from scratch using numpy or pure Python before moving to frameworks, and be sure to tackle the exercises in the textbooks. It’s crucial to solve problems and derive the math yourself instead of just reading.
I'll second that. These were the books I used at university. AIMA and Bishop in particular for fundamentals.
If you're looking to gain practical experience, you might want to consider building a project. For instance, you could create a generative AI that produces realistic images. Start with a specific goal and then refine it; for example, focus on improving image quality for certain contexts. Getting hands-on like this can teach you a ton! Of course, this is just one method; there are many ways to learn.
You might also want to check out Anthropic. They have some interesting resources that could fit your needs. Their website has various topics to explore, and you may find some courses that offer deeper insights into AI.
Since you already hold a degree, try approaching your study of AI like you would a thesis. Use Google Scholar to find foundational and niche papers and pair them with appropriate toolkits and textbooks.
If you're looking for a structured learning path with a college-level rigor, MIT's open courseware is an excellent option. You can access graduate-level classes, including lectures and lab work, all for free!
Instead of searching specifically for AI courses, focus on machine learning courses. The term AI often leads to content about newer models and tools, while machine learning resources often emphasize the mathematical foundations you’re looking for.
Wow! Thank you for the detailed answer. This is just the type of resource I’m looking for.