This is quite common in scientific research. The typical algorithm I follow is to reframe the problem in the language of different fields and see whether there is a more useful way of tackling it in that framework. There are always some leaks in the abstraction/translation but often by reframing the problem you find a good-enough solution.