# Foundation Molecular Grammar: Multi-Modal Foundation Models Induce Interpretable Molecular Graph Languages

arXiv: https://arxiv.org/abs/2505.22948 · HTML: https://mlsanity.com/paper/2505.22948/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 May 2025.

## Built on by (most cited first, 6 of 6)

- [Autonomous Agents for Scientific Discovery: Orchestrating Scientists, Language, Code, and Physics](https://mlsanity.com/paper/2510.09901/) (2025)
- [Graph Diffusion Transformers are In-Context Molecular Designers](https://mlsanity.com/paper/2510.08744/) (2025)
- [Post Hoc Regression Refinement via Pairwise Rankings](https://mlsanity.com/paper/2508.16495/) (2025)
- [Symbolic Neural Generation with Applications to Lead Discovery in Drug Design](https://mlsanity.com/paper/2510.23379/) (2025)
- [PolyFusionAgent: A Multimodal Foundation Model and Autonomous AI Assistant for Polymer Property Prediction and Inverse Design](https://mlsanity.com/paper/2605.26543/) (2026)
- [Higher-Order Molecular Grammars for Generative and Foundation Models in Chemistry](https://mlsanity.com/paper/2610.02186/) (2026)

## Abstract (arXiv)

Recent data-efficient molecular generation approaches exploit graph grammars to introduce interpretability into the generative models. However, grammar learning therein relies on expert annotation or unreliable heuristics for algorithmic inference. We propose Foundation Molecular Grammar (FMG), which leverages multi-modal foundation models (MMFMs) to induce an interpretable molecular language. By exploiting the chemical knowledge of an MMFM, FMG renders molecules as images, describes them as text, and aligns information across modalities using prompt learning. FMG can be used as a drop-in replacement for the prior grammar learning approaches in molecular generation and property prediction. We show that FMG not only excels in synthesizability, diversity, and data efficiency but also offers built-in chemical interpretability for automated molecular discovery workflows. Code is available at https://github.com/shiningsunnyday/induction.