Theory shows that huge amount of labelled data are needed in order to achieve reliable classification/regression methods when using deep/machine learning techniques. However, in the eye tracking field, manual annotation ...