
The author proposes a data-efficient and interpretable model for representing molecules using graph grammars, facilitating molecule generation and property prediction. The approach combines quality representation learning with the interpretability of rule-based grammar.


coremsg

representing-molecules-as-random-walks-over-interpretable-grammars-a-data-efficient-approach-for-molecular-discovery


Representing Molecules as Random Walks Over Interpretable Grammars: A Data-Efficient Approach for Molecular Discovery
