The development of electron and scanning probe microscopies in the second half of the twentieth century has produced spectacular images of the internal structure and composition of matter with nanometer, molecular, and atomic resolution. Largely, this progress was enabled by computer-assisted methods of microscope operation, data acquisition, and analysis. Advances in imaging technology in the beginning of the twenty-first century have opened the proverbial floodgates on the availability of high-veracity information on structure and functionality. From the hardware perspective, high-resolution imaging methods now routinely resolve atomic positions with approximately picometer precision, allowing for quantitative measurements of individual bond lengths and angles. Similarly, functional imaging often leads to multidimensional data sets containing partial or full information on properties of interest, acquired as a function of multiple parameters (time, temperature, or other external stimuli). Here, we review several recent applications of the big and deep data analysis methods to visualize, compress, and translate this multidimensional structural and functional data into physically and chemically relevant information.

中文翻译：

---

扫描和电子显微镜中的大数据和深度数据：从多维数据集导出功能。


二十世纪下半叶电子和扫描探针显微镜的发展已经产生了具有纳米、分子和原子分辨率的物质内部结构和成分的壮观图像。在很大程度上，这一进步是通过显微镜操作、数据采集和分析的计算机辅助方法实现的。二十一世纪初成像技术的进步为获得有关结构和功能的高精度信息打开了众所周知的闸门。从硬件角度来看，高分辨率成像方法现在通常以大约皮米的精度解析原子位置，从而可以定量测量各个键长和角度。类似地，功能成像通常会产生多维数据集，其中包含有关感兴趣属性的部分或全部信息，这些信息是作为多个参数（时间、温度或其他外部刺激）的函数而获取的。在这里，我们回顾了大数据和深度数据分析方法的一些最新应用，以可视化、压缩这种多维结构和功能数据并将其转换为物理和化学相关信息。