The most important algorithm to learn Boolean functions probably is the Fourier analysis based algorithm. To learn the algorithm, I think this paper is the best.

The most related paper to our Decision Diagram based Boolean learner is probably this paper and this thesis.

The field of computational learning theory is to study the complexity of learning. This field is part of study in Theoretical Computational Science. Hence, to understand the materials, you probably need some background in Computational Complexity.

A good book on Computational Complexity is this book from by Christos H. Papadimitriou. You can search the book at Amazon. I also found that materials in Randomized Algorithm and Communication Complexity can be useful. The Randomized Algorithm book by by Rajeev Motwani, Prabhakar Raghavan is a good book. The Communication Complexity book by Juraj Hromkovic describes many techniques that are fundamental to prove the complexity lower bounds. If you are really into theoretical CS study, you probably want to reference those books. If you only concerns on applications, you don't need to read those books. But it is always a good idea to keep in mind where to look for information if needed.

Our work in learning Boolean functions is heavily related to computational learning techniques for learning such as Deterministic Finite Automata (See lectures 23, 24 in the MIT lecture note). We are still in the process of understanding the limitation of Boolean learning and developing the Boolean learner.