To commemorate the significant role of the College of Nuclear Science as the pioneering field during the re-establishment of National Tsing Hua University (NTHU) in Taiwan, Academician Sow-Hsin Chen, a Distinguished Honorary Chair Professor of the College, has generously donated to his alma mater to co-establish the “Sow-Hsin Chen Distinguished Lectureship on Neutron Science and Technology.”
Professor Sow-Hsin Chen was a member of the first graduating class of the Institute of Nuclear Science, the inaugural graduate program established after NTHU was reestablished in Taiwan. Intended as a permanent and sustainable legacy, the Lectureship will be held every one to two years. It aims to invite internationally renowned experts and scholars to visit NTHU to deliver academic lectures on various theoretical and applied themes within neutron science. Through these exchanges of research expertise with local scholars, and by broadening the academic horizons of our young students’ academic horizons, the Lectureship seeks to solidify a robust foundation for NTHU’s cutting-edge research in neutron science and technology.

■ Academic events: 2026 Hideki Yukawa Nobel Prize Commemorative Series & Sow-Hsin Chen Honorary Lecture ■

Lecturer：Dr. Wei-Ren Chen
Title：From Structure to Inference: Perspectives on Neutron Scattering
Affiliation：Neutron Scattering Division/ Oak Ridge National Laboratory to Neutron Scattering Division, Oak Ridge National Laboratory
Research Field: Neutron Scattering Research of Soft Condensed Matter
Date： 2026.04.29 （WED.）13:30~15:00
Location：Learning Resource Center（學習資源中心）Macronix Building 旺宏館3F /Synchronized Distance-Learning Classroom-B（遠距教室B）

■ Abstract■

Neutron scattering has traditionally relied on regression-based fitting grounded in the formalisms of Fermi and Debye, where structural parameters are obtained by minimizing discrepancies between analytical models and experimentally measured intensities. In this lecture, I emphasize a conceptual transition from regression to inference in neutron scattering, a perspective shaped by my training under Professor Sow-Hsin Chen.

Neutron measurements are statistical sampling processes, and conventional least-squares regression can be understood as maximum likelihood estimation under specific noise assumptions. However, because scattering is an inverse problem with finite and noisy data, parameter fitting alone is insufficient. A Bayesian inference framework provides a more complete description by incorporating prior information, smoothness constraints, and uncertainty quantification into the analysis.

I discuss recent developments that extend classical regression into probabilistic structural determination, highlighting how inference-driven approaches improve robustness and interpretability. This movement from regression to inference represents a fundamental shift in how structural knowledge is extracted from neutron scattering experiments.