Enforcing feature decorrelation during training produces sharper saliency maps and higher accuracy on image classification benchmarks.
Saliency learning: Teaching the model where to pay attention
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SaliencyDecor: Enhancing Neural Network Interpretability through Feature Decorrelation
Enforcing feature decorrelation during training produces sharper saliency maps and higher accuracy on image classification benchmarks.