ALRITE learns two separate latent spaces, one for predicting treated counterfactuals and one for control counterfactuals, and combines them with an X-learner style weighting to estimate conditional average treatment effects.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Asymmetrical Latent Representation for Individual Treatment Effect Modeling
ALRITE learns two separate latent spaces, one for predicting treated counterfactuals and one for control counterfactuals, and combines them with an X-learner style weighting to estimate conditional average treatment effects.