CLP trains a perturbation network via meta-learning on causally augmented metadata to adjust classifier logits per sample, improving accuracy on four spurious-correlation benchmarks.
Learning imbalanced datasets with label- distribution-aware margin loss
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Mitigating Spurious Correlations with Causal Logit Perturbation
CLP trains a perturbation network via meta-learning on causally augmented metadata to adjust classifier logits per sample, improving accuracy on four spurious-correlation benchmarks.