Generative Chemical Language Models for Energetic Materials Discovery

TL;DR

This paper introduces generative chemical language models trained on large datasets, fine-tuned for energetic materials discovery, enabling accelerated design of high-performance compounds.

Contribution

It presents a transfer-learning framework applying molecular language models to energetic materials, extending beyond pharmacology and emphasizing fragment-based encodings.

Findings

Models can generate synthetically accessible energetic compounds.

Transfer learning improves model performance on specialized datasets.

Framework accelerates the discovery of next-generation energetic materials.

Abstract

The discovery of new energetic materials remains a pressing challenge hindered by limited availability of high-quality data. To address this, we have developed generative molecular language models that have been pretrained on extensive chemical data and then fine-tuned with curated energetic materials datasets. This transfer-learning strategy extends the chemical language model capabilities beyond the pharmacological space in which they have been predominantly developed, offering a framework applicable to other data-spare discovery problems. Furthermore, we discuss the benefits of fragment-based molecular encodings for chemical language models, in particular in constructing synthetically accessible structures. Together, these advances provide a foundation for accelerating the design of next-generation energetic materials with demanding performance requirements.