Physical theories are something difficult for mathematicians to understand, but they often produce interesting mathematical objects as their invariants (or observables in physical terms). Moreover it often happens that one physical theory gives rise to several mathematical objects, which look very different. However since they come from a same theory, there is a certain relation or duality between these different looking mathematical objects, and such duality can be an interesting subject of study in mathematics. In my talk I will present an instance of such duality arising from a recent study of four dimensional N=2 superconformal field theories.