Profile avatar
valenceai.bsky.social
Industrializing scientific discovery to radically improve lives. Powered by @recursionpharma.bsky.social
38 posts 340 followers 2 following
Regular Contributor
Active Commenter
comment in response to post
Members of the Valence team are also on the organizing committee for the LMRL workshop at #ICLR2025. Covered topics: multimodal representation learning, connecting molecular and biological data, modelling biological perturbations, and more. More details: lmrl.org
comment in response to post
We are also hiring for multiple roles across our offices in Montreal and London. If you’re interested in helping advance our mission of industrializing scientific discovery to radically improve lives, join us: www.valencelabs.com/careers
comment in response to post
Third place: “ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy” Our largest foundation model for cell microscopy to date Paper: arxiv.org/abs/2411.02572 Kian Kenyon-Dean
comment in response to post
Second place: ”GFlowNet Pretraining with Inexpensive Rewards” Atomic GFlowNets for de novo molecular design Paper: arxiv.org/abs/2409.09702 Mohit Pandey
comment in response to post
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
comment in response to post
5️⃣ "SynFlowNet: Design of Diverse and Novel Molecules with Synthesis Constraints" Where: Machine Learning in Structural Biology workshop in East Meeting Room Rooms 11 & 12 arxiv.org/abs/2405.01155
comment in response to post
4⃣ “ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy” Where: Foundation Models for Science workshop in West Meeting Room 202-204 arxiv.org/abs/2411.02572
comment in response to post
3⃣ “Towards Scientific Discovery with Dictionary Learning: Extracting Biological Concepts from Microscopy Foundation Models” Where: Interpretable AI workshop in East Ballroom A, B neurips.cc/virtual/2024...
comment in response to post
2⃣ “Score-Based Interaction Testing in Pairwise Experiments” Where: Causal representation learning workshop in East Exhibition Hall C neurips.cc/virtual/2024...
comment in response to post
1⃣ “Benchmarking Transcriptomics Foundation Models for Perturbation Analysis” Where: AIDrugX workshop in West Meeting Room 109, 110 arxiv.org/abs/2410.13956
comment in response to post
This year, we’re following this work with “SAFE setup for generative molecular design”. Come and find us tomorrow at the AI4Mat workshop at 12 PM PT in the West Meeting Room (211-214). Read the paper: arxiv.org/abs/2410.20232
comment in response to post
“ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy” Where: Foundation Models for Science workshop in West Meeting Room 202-204 When: Dec 15th at 8:15 AM PST arxiv.org/abs/2411.02572
comment in response to post
“Towards Scientific Discovery with Dictionary Learning: Extracting Biological Concepts from Microscopy Foundation Models” Where: Interpretable AI workshop in East Ballroom A, B When: Dec 15th at 8:15 AM PST neurips.cc/virtual/2024...
comment in response to post
“Score-Based Interaction Testing in Pairwise Experiments” Where: Causal Representation Learning workshop in East Exhibition Hall C When: Dec 15th at 8:15 AM PST neurips.cc/virtual/2024...
comment in response to post
“Benchmarking Transcriptomics Foundation Models for Perturbation Analysis” Where: AIDrugX workshop in West Meeting Room 109, 110 When: Dec 15th at 8:15 AM PST arxiv.org/abs/2410.13956
comment in response to post
“Graph Classification Gaussian Processes via Hodgelet Spectral Features” Where: Bayesian Decision-Making and Uncertainty workshop When: Dec 14th at 8:15 AM PST arxiv.org/abs/2410.10546
comment in response to post
“SAFE setup for generative molecular design” Where: AI4Mat workshop in West Meeting Room 211-214 When: Dec 14th at 8:15 AM PST arxiv.org/abs/2410.20232 A follow-up from our work last year: youtu.be/oJsj5vWmD3c
comment in response to post
“On the Scalability of GNNs for Molecular Graphs” Where: East Exhibit Hall A-C #3103 When: Dec 13th from 11 AM - 2 PM PST neurips.cc/virtual/2024... valencelabs.com/blog-posts/m...
comment in response to post
“Amortizing intractable inference in diffusion models for vision, language, and control” Where: West Ballroom A-D #7101 When: Dec 12th from 4:30 - 7:30 PM PT neurips.cc/virtual/2024...
comment in response to post
"QGFN: Controllable Greediness with Action Values” Where: West Ballroom A-D #6402 When: Dec 12th from 11 - 2 PM PST neurips.cc/virtual/2024...
comment in response to post
“Targeted Sequential Indirect Experiment Design” Where: West Ballroom A-D #7106 When: Dec 12th from 11 - 2 PM PST neurips.cc/virtual/2024...
comment in response to post
“Propensity Score Alignment of Unpaired Multimodal Data” Where: East Exhibit Hall A-C #1000 When: Dec 12th from 11 - 2 PM PST neurips.cc/virtual/2024... youtu.be/7a29mGz9LXI
comment in response to post
“How Molecules Impact Cells: Unlocking Contrastive PhenoMolecular Retrieval” Where: East Exhibit Hall A-C #1110 When: Dec 11th from 11 - 2 PM PST portal.valencelabs.com/blogs/post/m... neurips.cc/virtual/2024...
comment in response to post
“ET-Flow: Equivariant Flow Matching for Molecular Conformer Generation” Where: East Exhibit Hall A-C #2509 When: Dec 11th from 11 - 2 PM PST neurips.cc/virtual/2024...
comment in response to post
🧵 We hope OpenQDC will help advance the field of MLIP development toward a future of universal potentials with greater generalizability and robustness. Explore the library: www.openqdc.io Find us on GitHub: github.com/valence-labs...
comment in response to post
🧵 With OpenQDC, we standardized each dataset into a unified format to ensure fast loading, indexing and batching. We also compute missing values and provide additional labels while aggregating everything for access through a simple Pythonic API.
comment in response to post
🧵QM data can be challenging to work with. Datasets are stored in different formats with missing info (e.g. energy, distance, force units) needed for training machine learning interatomic potentials (MLIPs). Resources are also fragmented across code repositories and websites.