maybe in an ideal conditions three different forms could exists....but currently I think that the machine learning has the biggest chances... and should be some kind of fork from the crossvalidated. If everything goes well ( after several months or years) then NLP / Robotics / AI. can be split from this. but to have three proposal at the same time competing on the same audience.....could be a dead-end to all of them. please see: Closed AI project
Proposal: Artificial Intelligence & Robotics Proposal: Machine Learning Proposal: Natural Language Processing and Computational Linguistics (NLP/CL)