摘要
一位新闻记者-机器人与机器学习的工作人员新闻编辑每日新闻-机器学习的新研究是一篇报告的主题。根据NewsRx编辑在Penns Ylvania大学公园的新闻报道，研究表明：“材料表征仍然是一个劳动密集型的过程，需要大量的专家时间来后期处理和分析显微照片。因此，机器学习已经成为材料科学中不可或缺的工具，包括材料表征。”这项研究的财政支持来自国家科学基金会(NSF)。我们的新闻记者从宾夕法尼亚州立大学（宾州州立大学）的研究中得到一句话：“在这项研究中，”利用监督回归和分割模型对WSe 2薄膜原子力显微镜(AFM)高度图中晶体覆盖率的预测进行了深入的分析,从头开始训练回归模型s,从ImageNet和MicroNet上的ResNet预训练中通过转移学习来预测单层晶体覆盖率,从头开始训练模型d优于使用预训练模型s提取的特征。但微调产生了最好的性能，在不同的测试显微照片上获得了令人印象深刻的0.99r 2值。值得注意的是，从MicroNet中提取的特征明显优于从ImageNet中提取的特征，而在ImageNet上的微调则相反。由于该问题本质上是一个分割任务，因此该分割模型在确定图像块上的晶凹区域方面表现出色。当应用于全图像而不是斑块时，分割模型的性能明显下降，而回归系数则没有，这表明回归模型可能比分割模型对尺度和尺寸变化更稳健。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-New research on Machine Learning is th e subject of a report. According to news reporting out of University Park, Penns ylvania, by NewsRx editors, research stated, "Materials characterization remains a labor-intensive process, with a large amount of expert time required to post- process and analyze micrographs. As a result, machine learning has become an ess ential tool in materials science, including for materials characterization." Financial support for this research came from National Science Foundation (NSF). Our news journalists obtained a quote from the research from Pennsylvania State University (Penn State), "In this study, we perform an in-depth analysis of the prediction of crystal coverage in WSe 2 thin film atomic force microscopy (AFM) height maps with supervised regression and segmentation models. Regression model s were trained from scratch and through transfer learning from a ResNet pretrain ed on ImageNet and MicroNet to predict monolayer crystal coverage. Models traine d from scratch outperformed those using features extracted from pretrained model s, but fine-tuning yielded the best performance, with an impressive 0.99 R 2 val ue on a diverse set of held-out test micrographs. Notably, features extracted fr om MicroNet showed significantly better performance than those from ImageNet, bu t fine-tuning on ImageNet demonstrated the reverse. As the problem is natively a segmentation task, the segmentation models excelled in determining crystal cove rage on image patches. However, when applied to full images rather than patches, the performance of segmentation models degraded considerably, while the regress ors did not, suggesting that regression models may be more robust to scale and d imension changes compared to segmentation models."
NETL
NSTL