Profile avatar
nolovedeeplearning.bsky.social
human being | assoc prof in #ML #AI #Edinburgh | PI of #APRIL | #reliable #probabilistic #models #tractable #generative #neuro #symbolic | heretical empiricist | he/him 👉 https://april-tools.github.io
449 posts 4,788 followers 400 following
Prolific Poster
Conversation Starter
comment in response to post
yes, this looks relevant bsky.app/profile/made...
comment in response to post
that's quite cool, looking forward to reading this, it looks like an important practical problem
comment in response to post
indeed, we need this for a robust evaluation! is this always better than a regex, and how to prevent the auxiliary LLM to hallucinate?
comment in response to post
thanks! this is a bit different (if I understood correctly), I don't want to do constrained generation (e.g., according to a grammar), but just retrieve a sentence in the middle (supposed to be at the end) of a long piece of text that contains lots of "reasoning"
comment in response to post
RFK Jr's aggressive ignorance and scientific ineptitude is horrifying. He fundamentally does not understand or accept the most basic empirical principles for determining scientific information. I hope American medical science can quickly recovery from his absurd & tragic tenure as Health Secretary.
comment in response to post
Shameless plug: We’re trying to make it a little bit easier to get recognition for open-source contributions in ML by platforming them this year at our #ICML2025 workshop: codeml-workshop.github.io/codeml2025/
comment in response to post
co-organized with: @javaloyml.bsky.social @zzhe.bsky.social @poorvagarg.bsky.social John Leland (search.asu.edu/profile/4769...) Christoph Staudt (scholar.google.com/citations?us...) @lennertds.bsky.social Lingyun Yao (scholar.google.com/citations?us...)
comment in response to post
"IT'S EDUCATIONAL! (x4)"
comment in response to post
AI hate or AI hype hate?
comment in response to post
RoMBa!
comment in response to post
cyberpunk is the term!
comment in response to post
follow him here @dswatson.bsky.social !
comment in response to post
very obscure, what's the context?
comment in response to post
scheduling it on the NeurIPS deadline days is a statement!
comment in response to post
@neuralnoise.com
comment in response to post
oh! this sounds like a lovely workshop! I totally missed it, where is it going to happen? HW or UoE?
comment in response to post
I'd love to see the data/evidence. It sounds very strange, as such a correlation does not account of the quality of the discussion.