Deep Learning Removing Bias in Multi-modal Classifiers: Regularization by Maximizing Functional Entropies (סקירה) מיכאל ארליכסון, 12/03/2021