Bottom-up learning of feedback in a categorization task | IEEE Conference Publication | IEEE Xplore