材料研究中的可解释机器学习

王冠杰, 刘盛咸, 周健, 孙志梅

Explainable Machine Learning in the Research of Materials Science

WANG Guanjie, LIU Shengxian, ZHOU Jian, SUN Zhimei

图8 类别激活映射图(CAM)在图片分类问题中的示意图[98]

Fig.8 Examples of hypercolumn representations for eight image classes: cluster (a), fiber (b), matrix (c), matrix-surface (d), hraphene sheets (e), soot particles (f), high-density particles (g), and polymer residuals (h); Schematic of class activation map (CAM) in image classification problem[98] (Conv Block—convolutional neural network unit. FC Block—fully connected unit) (i)