2 ms·
> what you mean by association here then. Mutual information > but in the "True DAG" I'm unconvinced there is a "true" DAG and at best I think there's "th
by godelski 2y ago
> what you mean by association here then.
Mutual information
> but in the "True DAG"
I'm unconvinced there is a "true" DAG and at best I think there's "the most reasonable DAG given our observations." For all practical purposes I think this won't be meaningfully differentiable in most cases, so I'm fine to work with that. Just want to make sure we're on the same page.
> But my point is, at least, that we dont have access to this True model.
Then we're in agreement, but it's turtles all the way down. Everything is a model and all models are wrong, right? We definitely have more useful models, but there is always a "truer" model.
Why I was pushing against your example is because I think it is important to distinguish lack of association because the data to form the association is missing or unavailable to us (which may be impossibly unavailable; and if we go deep enough, we will always hit this point) vs a lack of association because the two things are actually independent[0]. One can be found via better sampling where the other will never be found (unfortunately indistinguishable from impossibly unavailable information).
> as-if you have the true model
Which is exactly why I'm making the point. We never have (or even have access to!) the "true" model. Just better models. That's why I say it isn't about being right, but less wrong. Because one is something that's achievable. If you're going to point to one turtle, for this, I think you might as well point to the rest. But there's still things that aren't turtles.
[0] I'll concede the to an argument of "at some point" everything is associated tracing back in time. Though I'm not entirely convinced of this argument because meta information.