niksmather

You can convert to numpy using .to_numpy().

It's also got much better support for more complex array shapes (e.g. each row storing an array). At least it did last time I used pandas!

I would also say Axler is much better prep for higher level applied math, as well as pure. If you are interested in how the big ideas of linear algebra extend to things like Fourier analysis it's very helpful to see the more abstract explanation of vector spaces.

So cool! Really looking forward to playing around with this.

It was also a great little explainer of how these tilings get made. One question: how do you guarantee/prove that the tiling you get back from the grid is quasiperiodic?

Pesticides are bad for bees, but Varroa is too. Until Varroa arrived in Australia the bees there didn't suffer from colony collapse, despite high pesticide use.

That's true of all statistical models, it's not some magic property of neural networks.

Great stuff - I really hope you manage to scale this up!

I can see the mathematical results are interesting, I was more wondering if there was a practical utility to this TreeFlow thing they built.

Apologies if I didn't understand the paper, but why do you want to apply diffusion models to tabular datasets in the first place?

Do we think they'll be better than decision trees? Is there some tabular problem that can be handled by diffusion but not trees?

Do neural networks work better than other models? They can definitely model a wider class of problems than traditional ML models (images being the canonical example). However, I thought where a like for like comparison was possible they tend to worse than gradient boosting.