Conditional Average Treatment Effect Distillation for Hypothetical State-Intervention Targeting Under Budget Constraints

Key points

• Flexible causal-effect models can be hard to operationalize, while coarse subgroup policies may lose individualized effect information.1
• We study budget-constrained hypothetical state interventions in observational data without a logged historical decision policy.1
• For obesity, elevated glucose, and lifetime smoking-history states, causal-structure-informed pretreatment covariates are prespecified, with capacity capped at 70% of those observed in the adverse state.1
• We evaluate causal distillation where Causal Forest and Bayesian Causal Forest teachers estimate conditional average treatment effects (CATEs), which decision tree and random forest students approximate.1

Sources (1)

• [1]Conditional Average Treatment Effect Distillation for Hypothetical State-Intervention Targeting Under Budget Constraints
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 01:08 AM

  Personalized public-health targeting must capture heterogeneous intervention benefits while remaining interpretable under resource constraints.

  Flexible causal-effect models can be hard to operationalize, while coarse subgroup policies may lose individualized effect information.

Extractive summary: sentences quoted from the sources.