In this answer Yves de Cornulier mentioned a talk about the possible uses of persistent homology in geometric topology and group theory. Persistent homology is a tool from the area of topological data analysis, specifically designed to extract information from empirical data and used for various applications, ranging from changes in brain function under drugs to the study of fluid flows to the merging of maps of distinct scales, along with many others. So this definitely belongs to the realm of applied mathematics, and it being used in pure mathematics is very interesting.
It goes without saying that many applications inspired a lot of research in pure mathematics, both in order to establish the foundations for the tools used in applies mathematics and just for the sake of studying interesting objects that appear in such interactions. I am talking specifically about the applied tools themselves being used in research in pure mathematics.
Going back in time, we find that a lot of methods that were initially developed mainly for some specific application, such as celestial mechanics, the stereometry of wine barrels or heat transfer, became the standard tools in pure mathematics. Now it seems that the flow of methods and techniques is mostly one-way, from pure mathematics to the applied. But not completely one-way, hence the question:
What are the recent uses of the tools from applied mathematics to the problems in pure mathematics?
If one requires a more specific indication what does "recent" mean, let's say last 30 years (but I would be delighted to hear about older examples as well).