How physics and machine learning joined forces to win the Physics Nobel ‘24
—
The 2024 Nobel Prize in Physics was just announced on 2024–10–04, and this year’s award went to Prof. John Hopfield and Prof. Geoffrey Hinton. But it’s a bit different this time. Instead of the usual discoveries about the natural world, the prize was awarded for something a little more artificial:
“for foundational discoveries and inventions that enable machine learning with artificial neural networks.”
Two machine learning (ML) models were highlighted: the Hopfield network and the Boltzmann machine. This may surprise some, as ML can seem far removed from physics. However, many early foundational ML concepts were inspired by physical systems.
What was the significance of the award? I believe it was to highlight recent advanced in generative AI (in text/image/video generations), and to remind us that the foundations of these modern models are rooted in physics. More specifically, Hopfield networks and Boltzmann machines can be thought of as some of the original generative models, borrowing physics principles from natural…