Breakthrough AI model predicts heat movement in materials 1,000,000 times faster than non-AI methods

A new machine-learning framework could revolutionize the efficiency of energy generation systems by predicting heat movement through semiconductors and insulators with unprecedented speed and accuracy.

Approximately 70 percent of generated energy is lost as waste heat, on average, in global energy systems, and addressing this inefficiency has been a major historical challenge for electrical engineers. Doing so will allow each unit of generated electricity to better supply consumer energy demands, allowing us to conserve energy and possibly even dramatically reduce carbon emissions.

Solving this problem, however, hinges on understanding the thermal properties of materials, a complex task due to the behavior of phonons—the subatomic particles that carry heat. The phonon dispersion relation (PDR), which describes the relationship between energy and momentum of phonons within a material’s crystal structure, is especially difficult to model.

Now, a team led by engineers at the Massachusetts Institute of Technology (MIT), has tackled this challenge with a new machine-learning framework that predicts PDRs up to 1,000 times faster than existing AI techniques and up to 1 million times faster than traditional methods.

This technique, described in a new paper published in the journal Nature Computational Science , promises more efficient power generation systems and microelectronics designs where heat management has traditionally been a significant bottleneck.

“Phonons are the culprit for thermal loss, yet obtaining their properties is notoriously challenging, either computationally or experimentally,” Mingda Li, an associate professor of nuclear science and engineering at MIT and the senior author of the paper, said in an MIT News release .

Heat-carrying phonons are difficult to predict due to their wide frequency range and variable travel speeds. Traditional machine-learning models, such as graph neural networks (GNNs), struggle with the high-dimensional nature of phonon dispersion relations. To overcome this, the researchers developed a virtual node graph neural network (VGNN), which introduces flexible virtual nodes to the fixed crystal structure, enabling the model to adapt and efficiently predict phonon behaviors.

“The way we do this is very efficient in coding. You just generate a few more nodes in your GNN. The physical location doesn’t matter, and the real nodes don’t even know the virtual nodes are there,” Abhijatmedhi Chotrattanapituk, a graduate student at MIT and co-author of the paper, said.

The VGNN can rapidly estimate phonon dispersion relations and offers slightly greater accuracy in predicting a material’s heat capacity, the study argued. This efficiency allows for the calculation of phonon dispersion relations for thousands of materials within seconds on a personal computer, potentially accelerating the discovery of materials with superior thermal properties.

Potential applications of new thermal-superior materials

Looking ahead, the researchers aim to refine their technique, enhancing virtual nodes’ sensitivity to capture minute changes affecting phonon structures.

“Graph nodes can be anything,” Li said. “And virtual nodes are a very generic approach you could use to predict a lot of high-dimensional quantities.”

This innovative framework not only promises to enhance energy efficiency but also opens new avenues in the study of optical and magnetic properties, potentially transforming multiple fields of material science.