(1/2)^5 ?

🤩 opensource.nibr.com/bamdd/

Perfect, thanks

Paywalled, unfortunately 😕

Wait until you find out that the mentioned „S-learner“ is what epidemiologists call „G computation“ for decades 😆

Also I‘m not sure if Nature Medicine really is a medical journal 😆

I don’t disagree but if you live in the ML world this might not be unusual. See this box from @alxndrmlk.bsky.social textbook. Note that Alexander abbreviates ITE correctly, using it for the -ized effect is stupid.

I don’t know the reference you are referring to but ML causal researchers often use a different and confusing language than epidemiologists and statisticians, for example, individualIZED treatment effects or heterogeneous treatment effects are the CATE.

Interesting, is the material available somewhere?

In defense of this paper, they mention individualized treatment effects, ie CATE, not ITE