An Early Investigation of Infants Learning
This research analyses and interprets psychological evidences of occlusion to investigate how an object could be represented and learnt that gives rise to infants' responses to the immediate environment from a computational point of view. Possible visual inputs are synthesized to construct a basic representation of an object that has not been addressed in the computational models reviewed. Followed by further studies of depth, mother recognition and cognitive development of other species, the basic representation construction connects with a straightforward process with desirable behaviours as output. Finally, tracking process, storage structure, representational input and process of responses are proposed. Future work includes development of an algorithm and implementation for verification.