A Expanded related works

In this section, we further expand on the related works described in Section 2. Moreover, all of these methods learn potential outcome functions that are shared between the different domains. A different strand of work aims to address the problem of identifiability of causal effects for an observational dataset of interest by leveraging data from other datasets. The paper assesses under which conditions such average causal effects can be transported according to the differences between the randomized and observational data. In this paper, we consider a shared label space.