The setup there is also a special case of our setup, where the reward is linear in the treatment vector, i.e. Hence, we omitted such an analysis. We will add some more elaborate discussion on these rates expanding on Remark 4. See response also to Reviewer #4 regarding efficiency. These extra moments can be used to construct more efficient estimators. However, these extra moment conditions have no bite in the case of homoskedastic noise.

To all reviewers, thank you very much for your thoughtful comments and suggestions. R#1: "...importance of similarity among the selected tasks... " R#1: "...domain randomization, when enough samples are used, is a better alternative to meta-learning... " R#2: "...Theorems 1 and 2 are asymptotic... " Hence, the theorems are NOT asymptotic. We will remove the asymptotic parts for clarity. R#2: 'Assumption 2 ... the per-task optimal models are centered around the corresponding optimal solutions. This assumption can easily be dropped with the cost of including the distance as a term.

To all reviewers, thank you very much for your thoughtful comments and suggestions. R#1: "...importance of similarity among the selected tasks... " R#1: "...domain randomization, when enough samples are used, is a better alternative to meta-learning... " R#2: "...Theorems 1 and 2 are asymptotic... " Hence, the theorems are NOT asymptotic. We will remove the asymptotic parts for clarity. R#2: 'Assumption 2 ... the per-task optimal models are centered around the corresponding optimal solutions. This assumption can easily be dropped with the cost of including the distance as a term.