.@Plinz What do you think of Judea Pearl’s causal inference and do-calc?
-
-
Replying to @DKedmey
I expect that causal dependency analysis, statistics and function approximation in AI will turn out to converge in a general theory of modeling.
2 replies 0 retweets 3 likes -
I think they will do very well for modeling static scenes, or for very distant dependencies with a long reaction time (so a stable state can be found reliably), but will collapse for modeling temporal dependencies, because their assumptions about time are incorrect.
1 reply 0 retweets 0 likes -
One word: homeostasis. Causality is too fucked up in a linear narrative of time to effectively model homeostasis.
1 reply 0 retweets 0 likes -
Causality just means modeling the world in terms of separate interfacing systems that change each other’s evolution. It is a mode of description.
1 reply 0 retweets 0 likes -
No. It means projecting events into a direct acyclic graph. It means imposing hierarchy into who is the biggest daddy of all.
1 reply 0 retweets 0 likes
Not the other way around? Any system‘s description is cyclic, but few macroscopic things are.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.