Artificial Intelligence has so far been strictly connected to big data: AI’s ability to perform and obtain trustable behaviours and information relies on data size. H4R is currently developing an AI that needs the smallest possible amount of data and that bases its learning process on a “speculative learning process”. Its activation is triggered by objective reality and its capability to reconstruct something addresses objects AI is not totally aware of.
Another aspect of our R&D is the implementation of self-evolutive functions on adaptive robots. This merging enables robots to learn to solve problems by themselves, adapting learning methods, skills and models to new situations.