Microsoft Research Releases Skala 1.1: Advancing Density Functional Theory Towards Predictive Accuracy
Summary:Microsoft Research has released Skala 1.1, the latest version of its deep-learning exchange-correlation functional designed to enhance the accuracy of Density Functional Theory (DFT) in computational chemistry. Trained on 2.5 times more data than its predecessor, Skala 1.1 demonstrates significant improvements in key areas such as thermochemistry, reaction kinetics, and molecular structure prediction. The model is now integrated into leading electronic-structure software packages like CP2K, Psi4
Microsoft Research Releases Skala 1.1: Advancing Density Functional Theory Towards Predictive Accuracy
Microsoft Research has released Skala 1.1, the latest version of its deep-learning exchange-correlation functional designed to enhance the accuracy of Density Functional Theory (DFT) in computational chemistry. Here are the key highlights of this release:
-
Enhanced Accuracy: Skala 1.1, trained on 2.5 times more data than its predecessor, demonstrates significant improvements in key areas such as thermochemistry, reaction kinetics, and molecular structure prediction. It ranks first in 32 out of the 55 categories in the widely used GMTKN55 benchmark, showcasing its broad and consistent accuracy.
-
Ecosystem Expansion: Skala is now integrated into leading electronic-structure software packages like CP2K, Psi4, FHI-aims, ORCA, and VASP, bringing next-generation DFT accuracy to these platforms. Additionally, a living benchmark has been introduced to track Skala's computational performance, providing the community with a transparent and continuously updated reference.
-
Data-Driven Improvements: The improvements in Skala are driven by the expansion of the Microsoft Research Accurate Chemistry Collection (MSR-ACC), which now includes new categories such as electron affinities and noncovalent clusters. This expansion has increased both the size and diversity of the training data, enabling Skala to systematically improve with each generation.
-
Performance Optimization: Skala 1.1 achieves computational costs comparable to semi-local functionals on both GPU and CPU platforms, with efficiency maintained across different software packages and hardware configurations. This ensures that users can benefit from high performance without compromising accuracy.
Technical Highlights
-
Continuous Improvement Model: Skala adopts a unique continuous improvement approach, where each new release is designed to supersede the previous one. This ensures that Skala can systematically enhance its accuracy while maintaining computational efficiency.
-
Cross-Platform Integration: Skala's integration extends beyond the open-source community to include collaborations with commercial software developers. This broad integration strategy ensures that Skala can be used by a wider range of researchers and engineers, promoting its adoption in practical applications.
-
Introduction of Performance Benchmarks: By introducing a living benchmark, Microsoft provides Skala users with a transparent and reliable tool for performance evaluation. This not only helps users choose the right computational tools but also provides important reference data for future research.
Industry Impact and Developer Recommendations
The release of Skala 1.1 marks a significant milestone in the field of computational chemistry. Its high accuracy and wide range of applications make it an ideal choice for researchers and engineers. For developers, it is recommended to stay updated with Skala's continuous improvements and actively participate in its open-source community to fully leverage its technological advantages.
Conclusion
The release of Skala 1.1 not only demonstrates Microsoft's leadership in AI-driven computational chemistry but also provides a powerful tool for future scientific research and industrial applications. As Skala continues to evolve and expand, it will play an increasingly important role in advancing fields such as materials science, catalysis, energy technologies, and drug discovery.
— END —Source: Microsoft Research AI (2026-08-20)
Tags: #Microsoft Research #Skala #Density Functional Theory #Computational Chemistry #AI for Science
Editorial & Fact-Checking Note: This article is compiled from primary research, official release documentation, and source papers by the ZICQ Newsroom pipeline with automated entity verification and human editorial review. If you notice any technical inaccuracy, please submit a correction via our corrections policy or email our editorial desk directly.
Community Comments