In brief
The Joint Mathematics Meetings showcased new perspectives in math and AI, addressing challenges and innovations in the field.
In Seattle, from January 8 to 11, the world’s largest assembly of mathematicians converged for the Joint Mathematics Meetings, drawing 5,444 attendees and featuring 3,272 talks. This year's event deviated from its traditional format by centering discussions around the official theme, “Mathematics in the Age of A.I.” This theme was proposed by Bryna Kra, president of the American Mathematical Society, in collaboration with 16 partner organizations.Dr. Kra described the theme as a crucial moment for the mathematical community. “A.I. is pervasive in our lives; it's essential to contemplate its influence on our pedagogy, our students, and our research,” she explained in an interview. She encouraged her colleagues to ponder questions such as the implications of having A.I. as a co-author.One of the event's highlights was a keynote lecture by Yann LeCun, chief A.I. scientist at Meta, who addressed attendees on the topic of “Mathematical Obstacles on the Way to Human-Level A.I.” LeCun humorously criticized the current state of machine learning, quipping, “The current state of machine learning is that it sucks,” which elicited laughter from the audience. He stressed the need to move beyond generative large language models, advocating instead for the development of a “large-scale world model” that possesses the ability to reason and plan based on its understanding of the world’s dynamics. Despite his forward-looking suggestions, he acknowledged ongoing challenges, including some mathematically intractable problems that remain unsolved.The Joint Mathematics Meetings served not only as a hub for sharing current research but also as a pivotal platform for grappling with the evolving landscape of mathematics in relation to artificial intelligence.
