The authors argue that a causal transition matrix P(Ŷ|do(Y),X) for instance-dependent label noise is identifiable when a noise-sensitive component X2 of the input is recovered, and they train a framework that separates X into X1 and X2.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Learning Causal Transition Matrix for Instance-dependent Label Noise
The authors argue that a causal transition matrix P(Ŷ|do(Y),X) for instance-dependent label noise is identifiable when a noise-sensitive component X2 of the input is recovered, and they train a framework that separates X into X1 and X2.