Guided Diffusion for the Discovery of New Superconductors

2025/09/29 by Prakash, Pawan, Gibson, Jason B., Li, Zhongwei +13 · 2 citations
#Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #FOS: Physical sciences #Materials Science (cond-mat.mtrl-sci) #Superconductivity (cond-mat.supr-con)

paper · doi:10.48550/arxiv.2509.25186

Abstract

The inverse design of materials with specific desired properties, such as high-temperature superconductivity, represents a formidable challenge in materials science due to the vastness of chemical and structural space. We present a guided diffusion framework to accelerate the discovery of novel superconductors. A DiffCSP foundation model is pretrained on the Alexandria Database and fine-tuned on 7,183 superconductors with first principles derived labels. Employing classifier-free guidance, we sample 200,000 structures, which lead to 34,027 unique candidates. A multistage screening process that combines machine learning and density functional theory (DFT) calculations to assess stability and electronic properties, identifies 773 candidates with DFT-calculated T_c>5 K. Notably, our generative model demonstrates effective property-driven design. Our computational findings were validated against experimental synthesis and characterization performed as part of this work, which highlighted challenges in sparsely charted chemistries. This end-to-end workflow accelerates superconductor discovery while underscoring the challenge of predicting and synthesizing experimentally realizable materials.

Citations

• Developing a Complete AI-Accelerated Workflow for Superconductor Discovery
• Diffusion Models without Classifier-free Guidance
• Open Materials Generation with Stochastic Interpolants
• Guiding a Diffusion Model with a Bad Version of Itself
• Unlocking Guidance for Discrete State-Space Diffusion and Flow Models
• MatterGen: a generative model for inorganic materials design
• Crystal Structure Prediction by Joint Equivariant Diffusion
• Crystal Structure Generation with Autoregressive Large Language Modeling
• Stochastic Interpolants: A Unifying Framework for Flows and Diffusions
• T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Models
• Adding Conditional Control to Text-to-Image Diffusion Models
• Adding Conditional Control to Text-to-Image Diffusion Models
• Flow Matching for Generative Modeling
• Classifier-Free Diffusion Guidance
• A universal graph deep learning interatomic potential for the periodic table
• High-Resolution Image Synthesis with Latent Diffusion Models
• High-Resolution Image Synthesis with Latent Diffusion Models
• Crystal Diffusion Variational Autoencoder for Periodic Material Generation
• ILVR: Conditioning Method for Denoising Diffusion Probabilistic Models
• Machine learning of superconducting critical temperature from Eliashberg theory
• AFLOW-XtalFinder: a reliable choice to identify crystalline prototypes
• Denoising Diffusion Probabilistic Models
• Q<scp>uantum</scp> ESPRESSO toward the exascale
• Attention Is All You Need
• From ultrasoft pseudopotentials to the projector augmented-wave method
• Efficient iterative schemes for<i>ab initio</i>total-energy calculations using a plane-wave basis set
• Transition temperature of strong-coupled superconductors reanalyzed

Cited by

• Machine Learning-Guided Discovery of Kagome Superconductors YRu3B2 and LuRu3B2
• On Approaches to Building Surrogate ODE Models for Diffusion Bridges

Related

• A Straightforward Gradient-Based Approach for High-Tc Superconductor Design: Leveraging Domain Knowledge via Adaptive Constraints
• Developing a Complete AI-Accelerated Workflow for Superconductor Discovery
• Closed-loop machine learning for discovery of novel superconductors
• Machine learning using structural representations for discovery of high temperature superconductors
• InvDesFlow-AL: active learning-based workflow for inverse design of functional materials