3 ms·

I'd say it's even less than that. They seem to be summarizing the ways the problem of teaching an agent to do anything (including be generally intelligent) can

I'd say it's even less than that. They seem to be summarizing the ways the problem of teaching an agent to do anything (including be generally intelligent) can be formulated as a problem of maximizing a reward (hence the title). Another way to look at it is, if we had a good enough function (e.g. a universal approximator) it can be made to model any behavior using numerical optimization. Which I think isn't very surprising, but apparently there is some arguments about it.

randcraw 5y ago
In fact, "if we had a good enough function" == "if we had sufficient funding". This refrain will resonate mightily in the willing ears of US congressfolk who want re-election and would rather talk about something other than Trump. So welcome back to the future, and the $trillions the US spent on 20 years of space race and 50 years of cold war. The catchphrase that motivates the next 50 years of government/corporate funding will be... They've got a Terminator and we don't.