The purpose of this symposium is to explore the application of machine learning to characterization with an emphasis on microscopy methods such as electron microscopy, probe microscopy, optical microscopy, and other imaging techniques. A natural point at the nexus of ML and such methods is image analysis and processing. Modern imaging systems have the further capability to generate data cubes, where the third dimension is a kind of spectral information. Thus in seeking characteristic "signatures" during data analysis one is exploring a much richer terrain, not simply looking for features or patterns in pictures. In addition to faster, "deeper", and more powerful image processing, additional issues that will be addressed include a broader role for ML improving our current capabilities and even enabling new modes and techniques.