Active Label Correction Using Robust Parameter Update and Entropy Propagation

Kwang In Kim

Abstract:

Label noise is prevalent in real-world visual learning applica-
tions and correcting all label mistakes can be prohibitively costly. Train-
ing neural network classifiers on such noisy datasets may lead to signifi-
cant performance degeneration. Active label correction (ALC) attempts
to minimize the re-labeling costs by identifying examples for which pro-
viding correct labels will yield maximal performance improvements. Ex-
isting ALC approaches typically select the examples that the classifier
is least confident about (e.g. with the largest entropies). However, such
confidence estimates can be unreliable as the classifier itself is initially
trained on noisy data. Also, na ̈ıvely selecting a batch of low confidence
examples can result in redundant labeling of spatially adjacent exam-
ples. We present a new ALC algorithm that addresses these challenges.
Our algorithm robustly estimates label confidence values by regulating
the contributions of individual examples in the parameter update of the
network. Further, our algorithm avoids redundant labeling by promoting
diversity in batch selection through propagating the confidence of each
newly labeled example to the entire dataset. Experiments involving four
benchmark datasets and two types of label noise demonstrate that our
algorithm offers a significant improvement in re-labeling efficiency over
state-of-the-art ALC approaches.