[10p-N302-1~10] AI for Measurement Science: Toward New Measurement Paradigms Enabled by the Deep Integration of Information and Measurement Sciences

2:35 PM - 3:05 PM JST(5:35 AM - 6:05 AM UTC)

[10p-N302-4] AI-Driven Measurement Science Enabled by Foundation Models and Active Learning:
From Experimental Design to Knowledge Discovery

〇Eiryo Kawakami^1,2,3 (1.The Univ. of Osaka, 2.Chiba Univ., 3.RIKEN)

Keywords:

AI-Driven Measurement Science,Foundation Models,Active Learning and Experimental Design Optimization

Recent advances in artificial intelligence (AI) are transforming measurement science from merely analyzing acquired data to designing and optimizing measurements themselves. In this talk, using examples from life science and medicine, I will introduce approaches for understanding complex phenomena through foundation models trained on large-scale data, as well as experimental planning based on active learning and Bayesian optimization. I will discuss the potential of “AI-driven measurement science,” where informative measurement conditions are efficiently explored under limited experimental resources, enabling iterative cycles of data acquisition, hypothesis generation, and validation.