Category Theory for Optimization

This paper shows how concepts coming from category theory can help to improve the algorithms dealing with large set of data. Data structures can be modeled by functors that are related by natural transformations usable both to reduce data size or to shift an algorithm applicable to a particular data structure to an equivalent algorithm for another data structure, ie. results are the same but time required to get it can be different. As an illustration, the paper takes the example of queries on graph databases used by semantic web and big data communities.