Tag
This paper proposes Dose-AIPTB, a framework for estimating the individual probability of treatment benefit under discrete dose assignments using attention-based aggregation, outperforming kernel alternatives in numerical experiments.