About
Since its inception in 2017, the Machine Learning and the Physical Sciences (ML4PS) workshop has served as a unique gathering space for the growing community spearheading cross-cutting research topics at the intersection of machine learning (ML) and the physical sciences (PS). This includes the applications of ML to problems in the physical sciences (ML for PS) as well as developments in ML motivated by physical insights (PS for ML). The physical sciences are defined inclusively, including but not limited to physics, astronomy, cosmology, chemistry, biophysics, materials science, and Earth science.
The physical sciences 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), and simultaneously tools and insights from the physical sciences have been used to develop efficient ML models (e.g., diffusion models and physics-informed neural networks). The communities coalescing at this workshop are encouraged to pursue fresh, synergistic solutions to "big science" questions that can spark new approaches in ML.
This year's programming explores the evolving interplay between academia and industry in basic research. Invited talks and panel discussions emphasize the myriad foundational and translational connections between these domains. Furthermore, we ask: how can our community help sustain open, curiosity-driven research in physics and related sciences when traditional funding models are under strain? What role can industry play in supporting fundamental science, and how can the insights emerging from basic research in the physical sciences and ML help catalyze wider innovations in industry and beyond?
NeurIPS 2025
The Machine Learning and the Physical
Sciences 2025 workshop will be held on December 6 or 7, 2025 at the San Diego Convention
Center
in San Diego, Californa (USA) as a part of the 39th annual
conference on
Neural Information Processing Systems (NeurIPS). The workshop is planned to take
place in a hybrid format inclusive of virtual participation.