115學年材料系-第七週 專題研討課程 演講公告 (115年10月22日)

日期：民國 115年 10月 22日(四)
時間：下午 3:20-5:10
地點：成功大學成功校區三系館鋼構區(3F)共同教室 A1302演講廳
主講人: 阮氏芳玲 助理教授
服務機關:國立臺灣科技大學 材料科學與工程系助理教授

題目：Machine Learning Approaches for Advancing AFM-based Nanomechanical
Measurements
機器學習於原子力顯微鏡量測開發及奈米機械性質之應用

簡介:Atomic force microscopy (AFM) enables quantitative characterization of the
nanomechanical properties of engineering materials through force curve
analysis. However, measurement reliability is constrained by uncertainties
associated with tip geometry, experimental conditions, contact model selection,
mathematical fitting, and variability in force curve characteristics. Conventional
analysis is also time intensive and highly dependent on user expertise,
particularly when processing large datasets. This presentation discusses machine
learning (ML) strategies for improving the accuracy, efficiency, and automation
of AFM-based nanomechanical characterization. Classification approaches are
introduced for contact model selection and material identification, while
regression approaches are explored for predicting elastic modulus directly from
characteristic features extracted from force curves. These strategies aim to
reduce dependence on prior knowledge of tip geometry and conventional
contact model fitting, thereby streamlining data processing and interpretation.
Key sources of measurement uncertainty are discussed alongside practical
applications of ML for automated analysis and mechanical property prediction.
By integrating AFM, contact mechanics, and ML, these developments offer a
promising framework for more reliable and efficient quantitative
nanomechanical measurements, with potential to advance the characterization of
engineering materials at the nanoscale.

歡迎各位撥冗參加!