Using probabilistic machine learning for augmented inteligence. Research tailored towards unifying models which perform low-level tasks into higher level ones by averaging outcomes (in a Bayesian way) using probabilistic graphical models. For example the research takes the Virtually Augmented Robot (VAR )- which completes a series of tests (1,2,3)- containing specialised modelling frameworks to do individual tasks. BOBO is created as a new learning model which leverages the benefits of VAR's models individually, but now can complete all the tasks using a single model.