I'm trying not to dox myself so I can be more open on HN (though more concerns in modern era...). You can find some harsh words against some ML community practices in my history and I think it is easy to get misinterpreted as calling people dumb or confuse academic bashing from utility (I criticize LLMs and diffusion a lot because I like them, not the other way). So yes and no. But the lectures I have aren't recorded and public (zoom for my Uni. I'm ABD in my PhD). My lecture slides and programs should be publicly visible though, but I don't go into this with them because I've been specifically asked to not teach this way :/ In all fairness, our ML course only has Calc 1 as a pre-req and CS students aren't required to take Lin Alg (most do though, but first courses are never really that great ime) or differential equations. TBH to get into this stuff you kinda need some metric theory. If you actually poke through this paper you'll find that come up very quickly, and this is common in the optimal transport community. But I think if you get into metric theory a lot of this will make sense pretty quickly. So if you can, maybe start with Shao's Mathematical Statistics?