At the ICLR conference in Rio, Recursion, in collaboration with Valence Labs, presented two groundbreaking machine learning models that promise to accelerate drug discovery. The first, TxFM, is a transcriptomics model that achieves superior performance compared to models up to 100 times larger in data size. The second, MarS-FM, is a new class of generative models designed for molecular dynamics simulations, enabling more efficient and accurate prediction of molecular behaviors. These advances highlight Recursion's continued commitment to using AI to revolutionize the pharmaceutical industry.
"We had a fantastic time at ICLR sharing our latest machine learning breakthroughs," said a Recursion representative.
The papers for both models are now available for review on OpenReview and arXiv.