Guiding Generative Models to Uncover Diverse and Novel Crystals via Reinforcement Learning

TL;DR

This paper presents a reinforcement learning framework that guides generative models to discover diverse, novel, and thermodynamically stable crystalline materials, addressing the challenge of exploring underrepresented regions in the design space.

Contribution

It introduces a novel RL-guided approach integrating group relative policy optimization with multi-objective rewards for controlled, property-guided crystal generation.

Findings

Enhanced generation of diverse and novel crystals.

Maintains chemical validity while targeting specific properties.

Provides a modular framework for controllable inverse design.

Abstract

Discovering functional crystalline materials entails navigating an immense combinatorial design space. While recent advances in generative artificial intelligence have enabled the sampling of chemically plausible compositions and structures, a fundamental challenge remains: the objective misalignment between likelihood-based sampling in generative modelling and targeted focus on underexplored regions where novel compounds reside. Here, we introduce a reinforcement learning framework that guides latent denoising diffusion models toward diverse and novel, yet thermodynamically viable crystalline compounds. Our approach integrates group relative policy optimisation with verifiable, multi-objective rewards that jointly balance creativity, stability, and diversity. Beyond de novo generation, we demonstrate enhanced property-guided design that preserves chemical validity, while targeting desired functional properties. This approach establishes a modular foundation for controllable AI-driven inverse design that addresses the novelty-validity trade-off across scientific discovery applications of generative models.