(no subject)
Nov. 21st, 2004 08:56 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
guslJudea Pearl has pretty cool stuff on causality. Perhaps I should apply for a PhD there too? I should be getting started with my applications.
(At UCLA, I would be close to many LJ people.)
Btw, when I talk about a "logic of experimental design", what I have in mind is:
A system for relating possible experiments and causal probabilistic models.
The kind of things we should be able to talk about include:
* observables
* confounding
* causation, distinct from correlation
* controlled variables (Pearl's do(.))
* noise (can always be modelled as the effect of an unseen factor)
* probabilities
And it would be nice to be able to talk about theoretical entities. This way we could distinguish between measurement noise and randomness in the process we are actually observing (not in practice, but in the theory).
i.e. essentially the sort of methodological questions that real scientists discuss. (This is why call myself an AI person: I like to formalize things I can do, and then teach a computer to do them. This has something to do with me being into phenomenology (i.e. introspection) ).
"phenomonologist": my excuse for being ...(what do you call someone too interested in his own mind?). It's not my fault: exploring my consciousness is so interesting...
(At UCLA, I would be close to many LJ people.)
Btw, when I talk about a "logic of experimental design", what I have in mind is:
A system for relating possible experiments and causal probabilistic models.
The kind of things we should be able to talk about include:
* observables
* confounding
* causation, distinct from correlation
* controlled variables (Pearl's do(.))
* noise (can always be modelled as the effect of an unseen factor)
* probabilities
And it would be nice to be able to talk about theoretical entities. This way we could distinguish between measurement noise and randomness in the process we are actually observing (not in practice, but in the theory).
i.e. essentially the sort of methodological questions that real scientists discuss. (This is why call myself an AI person: I like to formalize things I can do, and then teach a computer to do them. This has something to do with me being into phenomenology (i.e. introspection) ).
"phenomonologist": my excuse for being ...(what do you call someone too interested in his own mind?). It's not my fault: exploring my consciousness is so interesting...