Senior Research Scientist @ DeepMind. I work on LLM post-training, focusing on improving training data. Prior to that, I worked on AlphaFold. Typos are my own.
Oxford, England
Joined June 2018
- Tweets152
- Following315
- Followers2.7K
- Likes526
Pinned Tweet
A year ago I asked: Is there more than Self-Attention and Deep Sets? - and got very insightful answers. 🙏 Now, Ed, Martin and I wrote up our own take on the various neural networks architectures for sets.
Have a look and tell us what you think! :)
➡️fabianfuchsml.github.io/lear… ☕️
I wrote a blog post trying to understand TrackStar, a gradient-based method tracing LLM predictions to influential training examples.
Mostly: take the main equation, read it from right to left, and poke at the pieces with a small MNIST toy example. 🙂
☕️ fabianfuchsml.github.io/trac…
Fabian Fuchs retweeted
Our Apple ML Research team in Barcelona is looking for a PhD intern! 🎓
Curiosity-driven research 🧠 with the goal to publish 📝
Topics: Confidence/uncertainty quantification and reliability of LLMs 🤖
Apple here: jobs.apple.com/en-gb/details…
Graphs , Sets, Universality
We put more work into this and are presenting it via the ICLR blogpost track (thanks to organisers and reviewers!). Have a read and let us know what you think:
iclr-blogposts.github.io/202…
better in light mode💡, dark mode🌙 messes with the latex a bit
📢 New blog post!
Realising an intricate connection between PNA (@GabriCorso @lukecavabarrett @dom_beaini @pl219_Cambridge) & the seminal work on set representations (Wagstaff @FabianFuchsML @martinengelcke @IngmarPosner @maosbot), Fabian and I join forces to attempt to explain!
Fabian Fuchs retweeted
Text-to-image diffusion models seem to have a good idea of geometry. Can we extract that geometry? Or maybe we can nudge these models to create large 3D consistent environments? Here's a blog summarizing some ideas in this space :)
akosiorek.github.io/geometry…
Fabian Fuchs retweeted
📢 New blog post!
Realising an intricate connection between PNA (@GabriCorso @lukecavabarrett @dom_beaini @pl219_Cambridge) & the seminal work on set representations (Wagstaff @FabianFuchsML @martinengelcke @IngmarPosner @maosbot), Fabian and I join forces to attempt to explain!
I have recently had a range of very insightful conversations with @PetarV_93 about graph neural networks, networks on sets, universality and how ideas have spread in the two communities. This is our write up, feedback welcome as always! :)
➡️fabianfuchsml.github.io/univ… ☕️
I have recently had a range of very insightful conversations with @PetarV_93 about graph neural networks, networks on sets, universality and how ideas have spread in the two communities. This is our write up, feedback welcome as always! :)
➡️fabianfuchsml.github.io/univ… ☕️
Fabian Fuchs retweeted
Anyone know of a department looking to hire faculty in the protein/genome+evolution+ML space? Also RNA biology (asking for a friend) 🙂🥼🧪
Fabian Fuchs retweeted
New blog post! Find out:
- what reconstructing masked images and our brains have in common,
- why reconstructing masked images is a good idea for learning representations,
- what makes a good mask and how to learn one
akosiorek.github.io/ml/2022/…
Graph neural networks often have to globally aggregate over all nodes. How we do this can have a significant impact on performance 🎯. After we recently finished a project on this, I wrote a blog post on this topic. Let me know what you think! :)
➡️fabianfuchsml.github.io/equi… ☕️
Fabian Fuchs retweeted
Molecule Generation in 3D with Equivariant Diffusion (arxiv.org/abs/2203.17003). Very happy to share this project (the last of my PhD woohoo 🥳) and a super nice collab with @vgsatorras @ClementVignac (equal contrib shared among three of us) and of course @wellingmax
Emmy Noether connected symmetries and conserved quantities in physics - how is this related to exploiting symmetries with neural networks? 🤔
I've tried to answer this question in a blog post (no background knowledge required!):
➡️fabianfuchsml.github.io/noet… ☕️
Amazing video! Fantastic book, too, I am glad it receives all this attention. Many of those ideas & concepts are very fundamental and so helpful to understand, regardless of which specific sub-field of machine learning one is in.
Epic special edition MLST on geometric deep learning! Been in the works since May! with @mmbronstein @PetarV_93 @TacoCohen @joanbruna @ykilcher @ecsquendor youtu.be/bIZB1hIJ4u8
Fabian Fuchs retweeted
Metacognition in AI? We'd love for you to join us at #NeurIPS2021 and tell us about your work. Closing soon... #ai #artificialintelligence
#ml #machinelearning #robotics #cogsci sites.google.com/view/metaco…
Fabian Fuchs retweeted
What startups are out there these days that need or make good use of the probabilistic ML toolkit? (thinking mostly: generative modeling, prob. inference, BO/active learning)
Slowly starting to explore the job market and need your help!
🙏 RT
Fabian Fuchs retweeted
Just to clarify. RoseTTAFold uses SE3 transformers (the core part of the structure module) and Nvidia just released a much faster version of this core part. The current code does not yet integrate these updates.
Fabian Fuchs retweeted
RoseTTAFold about to become 21X faster?
"NVIDIA just released an open-source optimized implementation that uses 9x less memory and is up to 21x faster than the baseline official implementation."
developer.nvidia.com/blog/ac…
Thank you for this great write-up about AF2!
journals.iucr.org/d/issues/2…
"The SE(3)-Transformer [...] is currently too computationally
expensive for protein-scale tasks"
➡️ To be honest, I initially thought so as well; but @minkbaek et al made it work! 😍
science.sciencemag.org/conte…
Particularly the proof in chapter 4 (entirely @EdWagstaff 's work) blows my mind🤯. He worked on this proof for an entire year and, imho, totally worth it. I love his talent for making abstract theory accessible for the reader. Blog post will follow 🙂
arxiv.org/abs/2107.01959 ☕️
