A very general framework for doing research in MLAI

1. Define a problem that has a sufficiently significant impact on the field. 
2. Solve the problem with a framework that makes good sense. 
3. Derive a computational approximation to the proposed framework. 
4. Prove theoretical guarantees of the proposed framework and algorithm if possible. 
5. Design an experimental prototype to demonstrate the unique strength of the proposed algorithm and indicate where it might fail. 
6. Discuss the critical findings about the significance and the novelty of the proposed algorithm, and where it might epically fail (this is where future research is needed).