Dominique Beaini
@dom-beaini.bsky.social
470 followers 320 following 44 posts
Research Lead at @valenceai.bsky.social, Adjunct prof at @UMontreal and @Mila_quebec, interested in AI, deep learning, drug discovery, graphs, geometry, biology, chemistry, physics, philosophy, and the origin of life 🧠🦠 Aims to become a crazy scientist 👨‍🔬
Posts Media Videos Starter Packs
Pinned
dom-beaini.bsky.social
Excited to be presenting MolPhenix today as an #oral presentation at the @neuripsconf.bsky.social workshop: Foundation Models for Science!

3 out the 6 oral presentations in that workshop are from @valenceai.bsky.social / @recursionpharma.bsky.social

Join us in West Meeting Room 202-204 at 12pm
dom-beaini.bsky.social
Very excited by this work! Simultaneously learn the DFT and sample the Boltzmann distribution by using two GPUs concurrently.

Amazing stuff!
majhas.bsky.social
(1/n)🚨Train a model solving DFT for any geometry with almost no training data
Introducing Self-Refining Training for Amortized DFT: a variational method that predicts ground-state solutions across geometries and generates its own training data!
📜 arxiv.org/abs/2506.01225
💻 github.com/majhas/self-...
dom-beaini.bsky.social
🤯 We can now directly interact with biology via this new foundational knowledge graph. A must read 🧵
valenceai.bsky.social
1/ Introducing TxPert: a new model that predicts transcriptional responses across diverse biological contexts

It’s designed to generalize across unseen single-gene perturbations, novel combinations of gene perturbations, and even new cell types 🧵

www.valencelabs.com/txpert-predi...
TxPert: Predicting Cellular Responses to Unseen Genetic Perturbations - Valence Labs
We introduce TxPert: a state-of-the-art model that leverages multiple biological knowledge networks to accurately predict transcriptional responses under OOD scenarios.
www.valencelabs.com
dom-beaini.bsky.social
Everyone’s chasing AGI - but what’s the #DrugDiscovery equivalent?

Meet the Virtual Cell: our bold roadmap to curing any disease by simulating the building block of human biology

Check out @valenceai.bsky.social vision of curing diseases with GPU-based scientific discovery
valenceai.bsky.social
1/ At Valence Labs, @recursionpharma.bsky.social's AI research engine, we’re focused on advancing drug discovery outcomes through cutting-edge computational methods

Today, we're excited to share our vision for building virtual cells, guided by the predict-explain-discover framework 🧵
dom-beaini.bsky.social
Submit a 1-page abstract of your Neurips work to present it to the MoML conference on June 18th in Montreal
valenceai.bsky.social
(1/3)The poster submission deadline for MoML 2025 has been extended to May 20th, 2025.

Don’t miss an opportunity to share your work at this years conference.

Submit here: portal.ml4dd.com/moml-2025-po...
dom-beaini.bsky.social
MoML conference is back on June 18th!
valenceai.bsky.social
(1/5) The Molecular Machine Learning Conference (MoML) is back @mila-quebec.bsky.social this June 18th.

Join researchers advancing machine learning, molecular modeling, and therapeutic design.

Tickets are free for students.

Register Today: portal.ml4dd.com/events/molec...
dom-beaini.bsky.social
We're currently looking for a PostDoc position at @umontreal.ca and @mila-quebec.bsky.social to build a multimodal bacterial representation for anti-biotic with biologists from brunlab.com using:
- Phenomics
- Transcriptiomics
- Structural bio
- Viability
- ADME

DM me for more info
Brun Lab – Université de Montréal
brunlab.com
Reposted by Dominique Beaini
gsp-workshop.bsky.social
We’re absolutely thrilled to have an exciting lineup of speakers this year covering a wide range of topics, including @pseudomanifold.topology.rocks @dom-beaini.bsky.social @mathildepapillon.bsky.social Dhananjay Bhaskar, Naoki Saito!
Reposted by Dominique Beaini
valenceai.bsky.social
First place: “How Molecules Impact Cells: Unlocking Contrastive PhenoMolecular Retrieval”

Predicting the effect of molecules on cells with MolPhenix

Paper: www.arxiv.org/abs/2409.08302
@dom-beaini.bsky.social
Reposted by Dominique Beaini
phil-fradkin.bsky.social
Thanks to the FM4Science workshop at #Neurips for recognizing MolPhenix as best paper!

We had so much fun working on this with Puria (co-first author), @karush17.bsky.social, Frederik and co-supervised by Maciej and @dom-beaini.bsky.social

arxiv.org/abs/2409.08302
@valenceai.bsky.social
dom-beaini.bsky.social
Thanks! What are you up to these days?
dom-beaini.bsky.social
dom-beaini.bsky.social
Excited to be presenting MolPhenix today as an #oral presentation at the @neuripsconf.bsky.social workshop: Foundation Models for Science!

3 out the 6 oral presentations in that workshop are from @valenceai.bsky.social / @recursionpharma.bsky.social

Join us in West Meeting Room 202-204 at 12pm
dom-beaini.bsky.social
Definitely a career highlight to be awarded a best paper award at @neuripsconf.bsky.social FM4Science 🤩🤩

But even more exciting is that the 2nd and 3rd awards are given to colleagues Kian and Mohit, a podium sweep from @valenceai.bsky.social / @recursionpharma.bsky.social 🥇🥈🥉
dom-beaini.bsky.social
Excited to be presenting MolPhenix today as an #oral presentation at the @neuripsconf.bsky.social workshop: Foundation Models for Science!

3 out the 6 oral presentations in that workshop are from @valenceai.bsky.social / @recursionpharma.bsky.social

Join us in West Meeting Room 202-204 at 12pm
dom-beaini.bsky.social
We're hiringggggg

Interested? Come find me at @neuripsconf.bsky.social :)
valencelabs.com/careers

#machinelearning #AI #drugdiscovery
Reposted by Dominique Beaini
recursionpharma.bsky.social
Special delivery from @neuripsconf.bsky.social! We just released our latest certified dataset, RxRx3-core, which includes labeled images of 735 genetic knockouts & 1,674 small-molecule perturbations. The related benchmark evaluates the zero-shot prediction of compound-gene activity. #NeurIPS
👇
polarishub.io
🛡️ New certified dataset and benchmark from
@recursionpharma.bsky.social!

Did you know that high-dimensional data like cell images (i.e. phenomics) can help us understand the relationship between compounds and genes? 🧵

Dataset: polarishub.io/datasets/rec...
Benchmark: polarishub.io/benchmarks/r...
dom-beaini.bsky.social
Huge crowd at the Molphenix paper 🤩🤩 presented by @phil-fradkin.bsky.social and Puria
recursionpharma.bsky.social
Follow us to our spotlight paper on MolPhenix at
@neuripsconf.bsky.social - the foundation model that can predict the effect of any given molecule & concentration pair on phenotypic cell assays and cell morphology -- presented by researchers at @valenceai.bsky.social. 🧵
dom-beaini.bsky.social
Great news from Recursion! Getting closer to industrialization of drug discovery
recursionpharma.bsky.social
REC-617 is a testament to the potential of #AI guided precision chemistry & design to deliver new #cancer treatments. Today, we presented the latest interim data for our monotherapy Phase 1 trial for REC-617, a potential best-in-class CDK7 inhibitor, in a live Update Call. 🧵
dom-beaini.bsky.social
And of course @valence_ai and @Polaris_HQ
dom-beaini.bsky.social
Super excited to be flying to @neuripsconf.bsky.social to present 3 awesome papers! Including an oral presentation 📣!

Check out MolGPS, MolPhenix, and ET-Flow.

You can find us Wednesday and Friday at the posters, and Sunday for the oral!

Thanks to an amazing team of co-authors in the thread 🧵
dom-beaini.bsky.social
My students @majhas.bsky.social and @jyoonlee.bsky.social are now on bluesky!

They're new in the field, but they're assuming BIG! You should follow them to not miss out on their work.

Already a @neuripsconf.bsky.social paper only a few months after they started.
jyoonlee.bsky.social
We’re excited to present ET-Flow at #NeurIPS 2024—an Equivariant Flow Matching model that combines simplicity, efficiency, and precision to set a new standard for 3D molecular conformer generation.
🔖Paper: arxiv.org/abs/2410.22388
🔗Github: github.com/shenoynikhil...
dom-beaini.bsky.social
You were already there of course! Can't have ML for drug discovery without a great benchmarking platform
dom-beaini.bsky.social
ET-Flow shows, once again, that equivariance is better than Transformer when physical precision matters!

come see us at @neuripsconf.bsky.social !!
dom-beaini.bsky.social
Check out this new competition from @polarishub.io

bsky.app/profile/pola...
polarishub.io
🦠 We’re excited to announce our first competition in partnership with @asapdiscovery.bsky.social and @omsf.io!

Test your skills across three sub-challenges revolving around SARS-CoV-2 and MERS-CoV Mpro🧵

Full details: polarishub.io/competitions
Blog: polarishub.io/blog/antivir...