An EP algorithm for learning highly interpretable classifiers | IEEE Conference Publication | IEEE Xplore