Posts

Weight initialisation is key to both semi-supervised learning and meta-learning. What is its relationship to Bayesian priors?