Any supervised encoder must retain sensitivity along label-correlated directions, unifying non-robust features, texture bias, corruption fragility, and the robustness-accuracy tradeoff, and this is measurable and partially repairable via a new diagnostic and training term.
Simi- larity of neural network representations revisited
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Supervised Learning Has a Necessary Geometric Blind Spot: Theory, Consequences, and Minimal Repair
Any supervised encoder must retain sensitivity along label-correlated directions, unifying non-robust features, texture bias, corruption fragility, and the robustness-accuracy tradeoff, and this is measurable and partially repairable via a new diagnostic and training term.