Monroe: A Molecular Foundation Model for In-Context Probabilistic Inference

Summary

Monroe is a new molecular foundation model (MFM) pretrained on over 81 million molecules from the PM6 quantum chemistry dataset. It incorporates innovations including an improved graph representation of stereochemistry, improved training losses (conformer denoising and embedding decorrelation), improved multi-task learning, and uses a prior-data-fitted network (TabPFN) for downstream in-context prediction. Across established Polaris benchmarks, Monroe matches or exceeds existing MFMs, and on activity cliff benchmarks it achieves significant improvements over prior methods. Ablation and transfer experiments show that the PFN-based downstream predictors also substantially improve two leading existing models, MiniMol and CheMeleon, yielding new state-of-the-art variants called MiniMol_PFN and CheMeleon_PFN, indicating the strategy generalizes beyond Monroe. Source code is available at github.com/blazejba/monroe.