Guiding Generative Models to Uncover Diverse and Novel Crystals via Reinforcement Learning
This work addresses the novelty-validity trade-off in AI-driven inverse design for materials science, offering a modular foundation for controllable discovery of functional crystalline materials.
The paper tackles the problem of discovering novel crystalline materials by addressing the misalignment between likelihood-based generative models and the need to explore underexplored regions, introducing a reinforcement learning framework that guides diffusion models to generate diverse, novel, and thermodynamically viable compounds.
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.