>>11324867Nowadays AI is just machine learning which is just optimisation of the right loss function, which is almost always convex except in deep learning where it never is. If you want an introduction to the mathematics behind this topic I'd recommend any book or course on Convex Analysis and optimisation theory which has been studied to death and is the backbone of modern data "science".
Then, if you want to go into meme AI fields you should check bayesian inference (first) and reinforcement learning (second). A good book on the first is "The Bayesian Choice" by C.Robert and for the second you can check "Reinforcement Learning" by R. Sutton. You don't need to read the whole thing.
These two fields are usually rigorous, well formalized and combine into what is called model based reinforcement learning which is the current state of the art in reinforcement learning but with only meme applications at the moment. It is promising however as it could make reinforcement learning actually efficient and not just a meme.
