A Deep Generative Model for Fragment-Based Molecule Generation
This addresses molecule generation for cheminformatics, offering an incremental improvement over existing language model-based methods by reducing invalid and duplicate outputs.
The paper tackled the problem of generating valid and unique molecules by developing a deep generative model that builds molecules fragment by fragment instead of atom by atom, achieving state-of-the-art performance comparable to graph-based approaches with improved validity and uniqueness rates.
Molecule generation is a challenging open problem in cheminformatics. Currently, deep generative approaches addressing the challenge belong to two broad categories, differing in how molecules are represented. One approach encodes molecular graphs as strings of text, and learns their corresponding character-based language model. Another, more expressive, approach operates directly on the molecular graph. In this work, we address two limitations of the former: generation of invalid and duplicate molecules. To improve validity rates, we develop a language model for small molecular substructures called fragments, loosely inspired by the well-known paradigm of Fragment-Based Drug Design. In other words, we generate molecules fragment by fragment, instead of atom by atom. To improve uniqueness rates, we present a frequency-based masking strategy that helps generate molecules with infrequent fragments. We show experimentally that our model largely outperforms other language model-based competitors, reaching state-of-the-art performances typical of graph-based approaches. Moreover, generated molecules display molecular properties similar to those in the training sample, even in absence of explicit task-specific supervision.