Multi-robot cooperative Visual Question Answering