DailyGlimpse

Recursion Unveils AI Advances in Drug Discovery at ICLR 2026

AI
May 1, 2026 · 3:13 AM

Recursion, in collaboration with Valence Labs, presented cutting-edge machine learning breakthroughs at the International Conference on Learning Representations (ICLR) in Rio de Janeiro. The team showcased two key models: TxFM, a state-of-the-art transcriptomics model that outperforms competitors up to 100 times larger in data size, and MarS-FM, a new class of generative models for molecular dynamics simulations.

"We are excited to share these innovations that promise to accelerate drug discovery through AI," said a Recursion spokesperson.

The TxFM paper is available on OpenReview, while the MarS-FM paper can be found on arXiv. The presentations highlight Recursion's commitment to leveraging AI for connecting datasets and patients, ultimately driving better cancer therapies.