Bias in CNNs is detectable from internal representations—latent space geometry, layer activations, and filter weights—using statistical tests or a trained classifier, validated on 127,000+ models.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Unraveling Machine Behavior by Multi-Level Bias Analysis and Detection: Methodology and Application to Computer Vision
Bias in CNNs is detectable from internal representations—latent space geometry, layer activations, and filter weights—using statistical tests or a trained classifier, validated on 127,000+ models.