THOR AI completed the task 400 times faster than traditional simulations while maintaining full accuracy.
What took scientists a century to solve, THOR AI accomplished in mere seconds.
Researchers from the University of New Mexico and Los Alamos National Laboratory have developed the THOR AI (Tensors for High-dimensional Object Representation) framework, capable of tackling the once-impossible task of calculating configurational integrals in statistical physics.
Configurational integrals are essential for predicting how different materials behave under varying pressures and temperatures.
“Traditionally, solving the configurational integral directly has been considered impossible because it often involves thousands of dimensions. Classical integration techniques would require computational times exceeding the age of the universe, even with modern computers,” said Dimiter Petsev, professor in the UNM Department of Chemical and Biological Engineering.
In a recent breakthrough, the THOR AI framework used tensor train cross interpolation to break massive, high-dimensional datasets into smaller, manageable pieces.
“Tensor network methods, however, offer a new standard of accuracy and efficiency against which other approaches can be benchmarked,” he added.
The researchers also developed a specialized version of the method capable of detecting key crystal symmetries in materials within seconds.
THOR AI was tested on various material systems, including metals like copper and noble gases under extreme pressures. According to findings published in Physical Review Materials, it performed 400 times faster than traditional simulations without compromising accuracy.
The framework combines tensor networks with modern machine learning models, enabling it to evaluate materials under a wide range of conditions.
Given its efficiency, researchers say THOR AI could become a powerful tool for solving complex problems in physics and chemistry.
“THOR AI opens the door to faster discoveries and a deeper understanding of materials,” said Duc Truong, Los Alamos scientist and lead author of the study.
About the Author
