One of the challenges in modeling cognitive events from electroencephalogram
(EEG) data is finding representations that are invariant to inter- and
intra-subject differences, as well as to inherent noise associated with such
data. Herein, we propose a novel approach for learning such r