AI for Science Summer School 2026 Held on August 25–27, 2026

The AI for Science Summer School was held on August 25–27, 2026, at RIKEN Wako Campus, with online participation via Zoom. Jointly organized by the AI as Science Team and the DEEP-IN Working Group, the three-day school brought together more than 40 participants for lectures and hands-on coding sessions exploring how artificial intelligence can support scientific research. The program focused on three practical questions concerning where AI can help address scientific problems, how to choose appropriate methods, and how to apply them effectively.

Masato Taki (Rikkyo University/RIKEN) introduced machine learning fundamentals, and Lingxiao Wang (RIKEN/The University of Tokyo) discussed generative modeling. Vinicius Massami Mikuni (Nagoya University) presented foundation models and parameter inference. Naoya Takeishi (The University of Tokyo) covered simulation-based inference. Akinori Tanaka (RIKEN) and Conglong Li (Google DeepMind) explored large language models and their applications to scientific discovery.

Hands-on tutorials complemented the lectures. Tutors included Yang-yang Tan (The University of Tokyo), Jinyang Li (KEK/RIKEN), Victor Kawasaki-Borruat (EPFL), and Jingzong Zhang (Fudan University). The sessions covered supervised and unsupervised learning, generative models, data generation, and AI-assisted research workflows. These tutorials connected the lecture material with practical implementation.

By combining conceptual introductions with practical exercises, the school offered participants an opportunity to explore how AI methods could be incorporated into their own research. Its emphasis on scientific questions connected the different topics, from interpreting data and generating samples to assisting the research process itself. We thank all lecturers, tutors, participants, and iTHEMS staff for their contributions and support.

Reported by Lingxiao Wang