Midpoint Regularization: From High Uncertainty Training Labels to Conservative Classification Decisions.

Published in: ECML/PKDD (2) (2021)

Keyphrases

classification decisions
training set
training examples
supervised learning
training algorithm
training data
wide range
training process
training phase
decision theory
early stopping
inherent uncertainty
labelled data
fully labeled
data sets
labeled images
probability theory
class labels
learning algorithm