“We’re really firm believers that by contributing to the community and constructing upon open-source data models, the entire community moves further, faster,” says Larry Zitnick, the lead researcher for the OMat project.
Zitnick says the newOMat24 model will top the Matbench Discovery leaderboard, which ranks the very best machine-learning models for materials science. Its data set may even be one among the largest available.
“Materials science is having a machine-learning revolution,” says Shyue Ping Ong, a professor of nanoengineering on the University of California, San Diego, who was not involved within the project.
Previously, scientists were limited to doing very accurate calculations of fabric properties on very small systems or doing less accurate calculations on very big systems, says Ong. The processes were laborious and expensive. Machine learning has bridged that gap, and AI models allow scientists to perform simulations on combos of any elements within the periodic table far more quickly and cheaply, he says.
Meta’s decision to make its data set openly available is more significant than the AI model itself, says Gábor Csányi, a professor of molecular modeling on the University of Cambridge, who was not involved within the work.
“That is in stark contrast to other large industry players akin to Google and Microsoft, which also recently published competitive-looking models which were trained on equally large but secret data sets,” Csányi says.
To create the OMat24 data set, Meta took an existing one called Alexandria and sampled materials from it. Then they ran various simulations and calculations of various atoms to scale it.
Meta’s data set has around 110 million data points, which is persistently larger than earlier ones. Others also don’t necessarily have high-quality data, says Ong.
