On 1/20/2014 2:44 PM, Russell Standish wrote:
On Mon, Jan 20, 2014 at 09:41:04AM -0800, meekerdb wrote:
But Russell seems to think that "specific reason" means some
"objective", i.e. publicly determinable reason. In general one's
utility function is private, subjective and not known to others or
maybe even to yourself.
Not at all - the utility function need not be publicly known in order
to model it via observation. All that is required is for the agent's
actions to be consistent with optimising the utility in order to build
an effective model of the agent's utility via induction.
This still leads to agent predictability,
In the limit of arbitrarily many interactions. But that would be true even if the agent
did make random decisions where that is the best strategy. An opponent might (eventually)
have a model of the random distribution, but the agent would never become predictable or
even deterministic.
Brent
and the agent is neverthless
still deterministic.
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/groups/opt_out.
