Around 1,500 Latin inscriptions are discovered every year, offering an invaluable view into the daily life of ancient Romans—and posing a daunting challenge for the historians tasked with interpreting them.

But a new artificial intelligence tool, partly developed by Google researchers, can now help Latin scholars piece together these puzzles from the past, according to a study published on Wednesday.

Inscriptions in Latin were commonplace across the Roman world, from laying out the decrees of emperors to graffiti on the city streets. One mosaic outside a home in the ancient city of Pompeii even warns: “Beware of the dog”

Hierarchical Reasoning Models (HRM) tackle complex reasoning tasks while being smaller, faster, and more data-efficient than large AI models.

In contemporary artificial intelligence, transformers are everywhere, changing the way we do everything from natural language processing to computer vision. People have rushed to play with GPT-4 and other AI text models built on top of Transformer architectures because the machines are capable of solving problems they previously couldn’t, riffing stories, code or poetry, creating images from sentences, even speaking like Turing test-worthy humans. But as artificial intelligence improves, researchers have found that a grid-based or sequential approach to data has increasingly stringent constraints. And there’s a new AI technology that holds the potential of unraveling the mysteries that our complex and interconnected world holds around us: Graph Neural Networks (GNNs).

How might AI change the way we advance human knowledge? Could it change how universities like Cambridge carry out one of their core functions: research? Could AI be a technological transformation unlike anything we’ve seen before?