While the 2024 Nobel Prize in Physics celebrated the profound impact of the Physical Sciences (PS) on Machine Learning (ML) circa the 1980s, the symbiotic relationship between these fields has only grown stronger and more exciting since the era of Hopfield networks and Boltzmann machines. Since its inception in 2017, the Machine Learning for the Physical Sciences (ML4PS) workshop has served as a critical gathering space for the community spearheading cross-cutting research topics at the convergence of the physical sciences and machine learning.
The physical sciences offer large, information-rich datasets that promise to unlock the fundamental mysteries of nature, with scales ranging from high-resolution telescopes peering into the depths of the cosmos to particle colliders characterizing the smallest building blocks of the Universe. These complex datasets pose challenging, high-profile questions that invite innovation and opportunities for cross-pollination between ML researchers and physical scientists. Recent years have seen a tremendous increase in cases where ML models are used for scientific inference and discovery (e.g., geometric deep learning, simulation-based inference). Simultaneously, tools and insights from the physical sciences have been used to develop efficient and innovative ML models (e.g., diffusion models and physics-informed neural networks). The communities coalescing at this workshop often create fresh approaches to grand scientific challenges that also yield new methodologies in ML.
"ML and the physical sciences" in the late 2020s has taken on a double meaning: ML is not only enabling new kinds of physics discoveries across experiment and theory, but also, on a meta-level, AI systems including LLMs and coding agents are inspiring some researchers to question the very scaffolding of the traditional research cycle. In the next decade, who will be doing physics research and what exactly will that look like? Which of the established norms of the field should be preserved, and which might benefit from a fresh perspective? How can researchers today across all career levels help craft a near future of our field that feels exciting and empowering, that makes room for the expansive possibilities of cutting-edge tools without sacrificing trust in the scientific process?
This year, our programming highlights the special contributions of theoretical physics to our contemporary understanding of ML. Further, we hope to spark broad community engagement in envisioning possible future landscapes of AI & physics that support the flourishing of science and scientists alike.
