Convolutional Neural networks and Deep Learning are hot topics in Computer Science. These techniques can be used to identify handwriting, recognize objects, and model language. However, the use of these techniques for humanities research is still limited. In this talk, Melvin will highlight how existing techniques can be used to study conceptual shifts in text and images. Moreover, Melvin will argue that computer scientists need to work together with humanities researchers to push innovations. This requires methodological flexibility and openness from both computer scientists and humanities scholars.