Diversity regularization in deep ensembles

Changjian Shui, Azadeh Sadat Mozafari, Jonathan Marek +2

Calibrating the confidence of supervised learning models is important for a variety of contexts where the certainty over predictions should be reliable. However, it has been report…