Every indexed document

world peace and proving interesting theorems as indirect

not with the “external environment” as typically conceived

yet may still appear commonsensically intelligent.

and only she knows if she’s achieved them. An externallyprovided

humans often pursue complexly constructed rewards

which is a probability distribution over the space of environments

meaning that the total amount of reward they return to any agent is bounded

Legg and Hutter consider the class of environments that are

using multimodal communication (including verbalization

the efficient pragmatic general intelligence

and one over the space of possible goals

” given in [Goe93a]. After assuming a prior distribution over the

the kind of general intelligence it captures best

which draws on ideas from algorithmic information

with varying degrees of mathematical rigor. Perhaps the most carefullywrought

others may feel no need for any such thing.

divergent definitions abound; Legg and Hutter

elegant formalism to capture all of its meanings. Even in

with all the imprecision and contextuality that this generally entails.

which in the context of the present agent model is a method for handling