TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Catgrad: A categorical deep learning compiler

77 pointsby remexre3 months ago

7 comments

incrudible3 months ago
Category theory always reminds me of this quip by Feynman:<p><a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=B-eh2SD54fM" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=B-eh2SD54fM</a><p>When I hear people talking in the jargon of category theory, I do not understand what they say, but I have a suspicion that it is something rather mundane that I <i>would</i> understand if they were using the more specific terms for the given context. I understand the appeal of generalization from a mathematical perspective, but from a practical programming perspective, I fail to understand the value proposition.
评论 #42965610 未加载
评论 #42968335 未加载
评论 #42963065 未加载
abeppu3 months ago
So, learning about the categorical structure is interesting, but is there a specific advantage to seeing these concepts directly informing the implementation vs as a post-hoc way of explaining what autodiff is doing? E.g. Tensorflow is creating and transforming graphs of computation. Despite being written before most of the categorical work cited was done, isn&#x27;t it doing the &quot;same&quot; thing, but we just wouldn&#x27;t find names or comments in the code that closely align with the categorical work?
评论 #42955985 未加载
tripplyons3 months ago
How does this compare to the XLA&#x27;s ability to compile full training steps from JAX?
catgary3 months ago
This is a nice project, but “You only linearize once” is more-or-less the type theory version of “Reverse Derivative Categories”, so JAX really does this already.
评论 #42955026 未加载
drumnerd3 months ago
I think about cones all the time when doing machine learning. If I have an object O that can be mapped to A and B I can learn a function from A to B or B to A if I can generate Os. That’s my view of self supervised learning.
bguberfain3 months ago
It remembers me Theano [0].<p>[0] <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Theano_(software)" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Theano_(software)</a>
cyanydeez3 months ago
i always find it fascinating when projects like these can seem to find a proper overview for the AI curious non-PhD students.