Monroe: A Molecular Foundation Model for In-Context Probabilistic Inference
Abstract
Bioassay activity prediction is often data-limited because drug-discovery datasets rely on time-consuming and expensive wet-lab experiments for data generation and evaluation. This challenge has inspired recent research into molecular foundation models (MFMs), which aim to encode general-purpose chemical knowledge into molecular representations that generalize well in data-constrained scenarios. This paper presents Monroe, a new MFM with several innovations over the existing state of the art: increased scale allowing pre-training on over 81 million molecules from the PM6 quantum chemistry dataset; improved graph representation of stereochemistry; improved training losses including conformer denoising and embedding decorrelation; improved multi-task learning; and the use of a prior-data-fitted model (TabPFNv3) for downstream in-context prediction. Our evaluations use a principled pairwise comparison framework that measures statistically significant performance differences. Across established Polaris benchmarks, Monroe matches or exceeds existing MFMs, while on activity cliff benchmarks, designed to assess utility for molecular discovery, it achieves significant improvements over prior methods. Finally, ablation and transfer experiments show that PFN-based downstream predictors also substantially improve two leading existing models, MiniMol and CheMeleon, yielding new state-of-the-art variants we call MiniMol_PFN and CheMeleon_PFN, suggesting that our downstream adaptation strategy generalizes beyond Monroe. Source code is at github.com/blazejba/Monroe
Related
- Bayesian Gaussian process inference of RNA velocity as a probabilistic vector field on single-cell manifolds
- Context-dependent protein structure prediction analysis and complex stoichiometry inference with MultimerMapper.
- Molecular Grammar of Intrinsically Disordered Regions: How Sequence and Context Shape Biomolecular Condensation and Disease
- How 'Foundational' Are Current Molecular Foundation Models?
- Chemical foundation models reveal molecular principles underlying glycobiology
BibTeX
@misc{banaszewski2026monroe,
title = {{Monroe: A Molecular Foundation Model for In-Context Probabilistic Inference}},
author = {Blazej Banaszewski and Andrew W. Fitzgibbon},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2608.18982},
eprint = {2608.18982},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.18982},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Monroe: A Molecular Foundation Model for In-Context Probabilistic Inference
AU - Blazej Banaszewski
AU - Andrew W. Fitzgibbon
PY - 2026
DA - 2026/11/06
PB - arXiv
DO - 10.48550/arxiv.2608.18982
UR - https://arxiv.org/abs/2608.18982
AB - Bioassay activity prediction is often data-limited because drug-discovery datasets rely on time-consuming and expensive wet-lab experiments for data generation and evaluation. This challenge has inspired recent research into molecular foundation models (MFMs), which aim to encode general-purpose chemical knowledge into molecular representations that generalize well in data-constrained scenarios. This paper presents Monroe, a new MFM with several innovations over the existing state of the art: increased scale allowing pre-training on over 81 million molecules from the PM6 quantum chemistry dataset; improved graph representation of stereochemistry; improved training losses including conformer denoising and embedding decorrelation; improved multi-task learning; and the use of a prior-data-fitted model (TabPFNv3) for downstream in-context prediction. Our evaluations use a principled pairwise comparison framework that measures statistically significant performance differences. Across established Polaris benchmarks, Monroe matches or exceeds existing MFMs, while on activity cliff benchmarks, designed to assess utility for molecular discovery, it achieves significant improvements over prior methods. Finally, ablation and transfer experiments show that PFN-based downstream predictors also substantially improve two leading existing models, MiniMol and CheMeleon, yielding new state-of-the-art variants we call MiniMol_PFN and CheMeleon_PFN, suggesting that our downstream adaptation strategy generalizes beyond Monroe. Source code is at github.com/blazejba/Monroe
ER -