Michael Bronstein is a leading expert in machine learning and artificial intelligence. Bronstein recently took up the post of Director of the new AITHYRA Institute of the Austrian Academy of Sciences. The institute was established with the support of the non-profit Boehringer Ingelheim Foundation, which is contributing 150 million euros to its funding.

Bronstein shared his thoughts on the groundbreaking recognition of AI in this year’s Nobel Prizes for Chemistry and Physics. For the first time in history, the Nobel committees awarded achievements related to AI-driven breakthroughs, signaling the transformative power of artificial intelligence in scientific research.

Nobel Prizes just a foretaste

In discussing the Chemistry Prize, Bronstein highlights the monumental work of Demis Hassabis and John Jumper from Google’s DeepMind as well as of David Baker, emphasizing their contributions to computational protein design. “It’s hard to think of individuals more deserving of a Nobel Prize than the DeepMind AlphaFold team and David Baker, a guru of computational protein design,” Bronstein remarks.

He notes how their work addressed one of the “holy grail” problems in structural biology, revolutionizing the field by enabling solutions to challenges previously considered unsolvable. This milestone is particularly exciting for Bronstein’s newly launched institute, AITHYRA, which focuses on the intersection of AI and biomedicine. “This week's Nobel Prizes, which for the first time recognize AI for science, give tremendous momentum to AITHYRA and are just a foretaste of what’s to come in this field.”

However, Bronstein also points out that the success of AlphaFold owes a significant debt to an often-overlooked contributor—the Protein Data Bank (PDB). This database, painstakingly built by hundreds of teams over nearly five decades, provided the structural protein data that was crucial in training AlphaFold. “Perhaps the unsung hero of the AlphaFold Nobel Prize is the Protein Data Bank (PDB),” he says, emphasizing the need to invest in similar foundational resources for future AI advances. “While many are focused on developing the next AlphaFold, we, as a community, should also consider how to create the next PDB. In my view, new biological data sources should be developed in parallel with the AI methods that will utilize them.”

“Did not expect it coming”

Michael Bronstein also reflects on the Nobel Prize in Physics, awarded to John Hopfield and Geoffrey Hinton, pioneers in the understanding and design of neural networks. He is convinced that deep learning has had a profound impact on nearly every field of science and technology. "In my opinion, deep learning is a Nobel-worthy invention and long overdue given the profound impact these methods have had," he states.

Bronstein also acknowledges the surprise of seeing AI-related work recognized in Physics: “Since there is no Nobel prize in computer science, I frankly did not expect it coming.” Still, Bronstein praised the Nobel committee for its decision, calling the recognition "very apt" and congratulating the laureates for their well-deserved achievement.

Note: This text was created with the help of AI.