In biology, unpaired data is the norm as experiments are often destructive. This limits our ability to build powerful multimodal models for drug discovery. To fuel a multimodal future in TechBio, we propose a simple algorithm for matching. 🧵 youtu.be/7a29mGz9LXI

🚀Introducing — Meta Flow Matching (MFM) 🚀 Imagine predicting patient-specific treatment responses for unseen cases or building generative models that adapt across different measures. MFM makes this a reality. 📰Paper: arxiv.org/abs/2408.14608 💻Code: github.com/lazaratan/meta…

Excited to share my internship project with Jason Hartford! Perturbation interaction testing on unstructure obs is challenging because: (1) without accessing the underlying mechanism, the notion of interaction is unclear; (2) there's no principled way to quantify interactivity...

This paper will appear at NeurIPS 🎉 Looking forward to seeing you all in Vancouver. Congrats to Johnny , ZuH_dav and Jana Osea.

Really happy to see my ValenceLabs internship project out! This was really cool to work on because it really highlighted a few things that I believe are required to build successful ML+Bio projects:

Reminder - applications for Ellis PhDs close in two weeks. I'll be accepting one student through the Ellis program this year so if you're interested in working together, consider applying.

Statistical ecologist Dr. Marie Auger-Méthé (Marie Auger-Méthé), of UBC Statistics & UBC Oceans (also found at ubcoceans.bsky.social), will be inducted as a Member of the Royal Society of Canada on Friday. RSC // SRC RSC // SRC

In addition to the ELLIS PhD (closing this week), I'm looking for two more PhDs funded through UKRI's AI CDT program: Multimodal representation learning w.@MagnusRattray findaphd.com/phds/project/u… Casual abstraction of dynamical systems w. Mauricio Alvarez findaphd.com/phds/project/c…

We're hiring two interns and a full time post in the Valence labs London office. If you're interested in causal inference / generative models / latent variable inference & how to impact real world drug discovery programs, consider applying: job-boards.greenhouse.io/valencelabs

Planning to attend the Superbowl of ML conferences? NeurIPS Conference is coming Dec. 10-15 to the Vancouver Convention Center and researchers from Valence Labs powered by Recursion will be on site presenting over 10 accepted papers & hosting a “Night of Science” #TechBio social

I have a poster Thursday morning on aligning unpaired perturbational multimodal data at #NeurIPS2024 East Exhibit Hall A-C #1000 Jason Hartford also has a PhD ad to follow up on this project: findaphd.com/phds/project/u… Come talk to us if it seems of interest! cc: ZuH_dav

🎉 It’s a clean sweep! This past weekend at NeurIPS Conference, the world's most prestigious #ML conference, papers from Valence Labs & Recursion won first, second and third place awards at the Foundation Models for Science workshop. The winning papers offer breakthroughs in leveraging

I had a blast working with such an amazing team! Marta Skreta Joey Bose Alex Tong Kirill Neklyudov Check out our paper for details and code for easy to use examples! 📄Paper: arxiv.org/abs/2412.17762 💻Code: github.com/necludov/sup... 🤗HuggingFace: huggingface.co/superdiff

Characterization of Allobodo yubaba sp. nov. and Novijibodo darinka gen. et sp. nov., cultivable free-living species of the phylogenetically enigmatic kinetoplastid taxon Allobodonidae Julia A. Packer et al. onlinelibrary.wiley.com/doi/10.1111/je…

Distinguishing Cause from Effect with Causal Velocity Models ift.tt/uBS7xmF

Distinguishing Cause from Effect with Causal Velocity Models. arxiv.org/abs/2502.05122

just spotted on the Singapore skyline, coming to an iclr near u tmrw... #superdiff #ICLR2025 🦹‍♀️ (Hall 3 + Hall 2B #141 @ 3 pm)