GEWUM: General Exploration Workflow for the Utopia of Materials: A Unified Platform for Automated Structure Generation, Selection, and Validation

Abstract

The fragmented landscape of existing computational tools often hinders the seamless integration of large-scale structure prediction with rigorous stability validation. To address this, we present GEWUM (General Exploration Workflow for the Utopia of Materials), an open-source platform that integrates the Selective Random Structure Search (SRSS) strategy with universal Machine Learning Interatomic Potentials (uMLIPs) to automate and accelerate materials discovery. With native support for SLURM-based HPC clusters, GEWUM unifies the entire workflow from structure generation and diversity-preserving selection to thermodynamic/dynamic stability assessments and property calculations. The platform further incorporates built-in visualization tools, including Sankey diagrams for space-group evolution, violin plots for energy distributions, and t-SNE/UMAP embeddings for structural diversity, enabling intuitive interpretation of screening results. We demonstrate GEWUM through three case studies: low-energy polymorph prediction in Al-Sc-N, identification of a P-62c phase of U3Si5, and high-pressure structure prediction of ThH10 at 150 GPa. Benchmark tests confirm reasonable agreement in thermophysical property predictions.

References in corpus (14)

• CALYPSO: a method for crystal structure prediction
• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• Implementation strategies in phonopy and phono3py
• Ab initio Random Structure Searching
• A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
• AlScN: A III-V semiconductor based ferroelectric
• Data-driven materials science: status, challenges and perspectives
• AiiDA: Automated Interactive Infrastructure and Database for Computational Science
• On-the-fly Closed-loop Autonomous Materials Discovery via Bayesian Active Learning
• Origin of the anomalous piezoelectric response in wurtzite Sc${}_x$Al${}_{1-x}$N alloys
• Superconductivity at 161 K in Thorium Hydride $Th{H}_{10}$: Synthesis and Properties
• A critical examination of compound stability predictions from machine-learned formation energies
• PyXtal: a Python Library for Crystal Structure Generation and Symmetry Analysis
• aflow++: a C++ framework for autonomous materials design