Information theory provides a nice way to quantify this, but it's not the only possible avenue
Show this thread
Intelligence is the degree of efficiency with which a learning system turns experience (data) into generalizable programs. I = G / E In AI, "system" should be understood as including the human engineers. Most of the "data -> generalization" conversion happens during model design
-
-
-
A proper mathematical theory of abstraction should provide formal ways to quantify generalization.
Show this thread
End of conversation
New conversation -
-
-
It is much better to call such a quantity other than intelligence. I could not focus on what you said because you said intelligence. Why not, say, generalization-to-data ratio (GDR) as if signal-to-noise ratio?
-
So agreed. Seems to me as well that the proposed ratio talks about Machine Learning + data engineers ability, but not about Cognition, so real intelligence.
End of conversation
New conversation -
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This Tweet is unavailable.
-
I think he means it’s about the degree of efficiency....
End of conversation
-
-
-
This Tweet is unavailable.
- Show replies
New conversation -
-
This is helpful. Creating even *relevant models can daunting. At what point in the design phase specifically `G`, can you recognize or offer preliminary indicators as to what what would be more efficient..
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.