All machine learning is just renormalization which in turn is a convolution in H... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		adamnemecek on Oct 8, 2023 \| parent \| context \| favorite \| on: Decomposing language models into understandable co... All machine learning is just renormalization which in turn is a convolution in Hopf algebra. That's why you see superposition "In physics, wherever there is a linear system with a "superposition principle", a convolution operation makes an appearance." I'm working this out in more details but it is uncanny how much it works out. I have a discord if you want to discuss this further https://discord.cofunctional.ai

esafak on Oct 8, 2023 | [–]

Do you mean all ML or just large neural networks? Where is renormalization in a tree model? What superposition are you referring to?

adamnemecek on Oct 8, 2023 | | [–]

Renormalization is all about this symmetric partitioning.

LeonigMig on Oct 8, 2023 | [–]

I suppose we should be cautious, the human mind is capable of overfitting too

adamnemecek on Oct 8, 2023 | [–]

You have no clue what you are talking about.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact