Glen writes:
< So we might be able to measure the "emotion" of a chatbot by observing
something like "stress" in the execution of the algorithm. If, for example,
there's a recursive "subroutine" inside it, a deep iteration of that
"subroutine" might indicate a more "stressful" prompt, whereas a shallow
iteration might indicate "relaxation" or "comfort". >
A fear response could be resolved by having a coherent way to identify and sort
risks and to understand likely outcomes and mitigations. In absence of these,
attention could be focused on fear itself and there could be feedback
amplification. An algorithm that prioritizes inputs and acts on them when
present would save energy so long as it could yield (idle) when a match for
risk stimuli was not identified. On the other hand, a tight
"while(TERRIFIED) { worry(); }" would run a CPU hot. In this view stress
having physical correlates (energy, heat) depends on the available cognitive
tools.
Same outrage. I just saw Val Demings expressing outrage about gun violence in
her debate with Marco Rubio. Did she mean it? Did she rehearse how to seem
more emotional than she really was? Perhaps it isn't even knowable without
much pre-measurement, e.g., fMRI prior to the expression. It could be she
taught herself how to spin her own emotions up, and in the moment, they were
"real" emotions even though they were choregraphed in a rational way. When it
comes to simulation or generative learning, there is the natural latency scale
for humans of say 1/10th of a second, and so long as a calculation could beat
that computational deadline, then it could insert whatever dramatic "rest
notes" that were needed to appear physiologically plausible.
Marcus
-. --- - / ...- .- .-.. .. -.. / -- --- .-. ... . / -.-. --- -.. .
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom
https://bit.ly/virtualfriam
to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com
FRIAM-COMIC http://friam-comic.blogspot.com/
archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/
1/2003 thru 6/2021 http://friam.383.s1.nabble.com/
