[ExI] i got friends in low places

[spike at rainier66.com]

-----Original Message-----
From: extropy-chat <extropy-chat-bounces at lists.extropy.org> On Behalf Of
BillK via extropy-chat
Subject: Re: [ExI] i got friends in low places

On Thu, 25 Aug 2022 at 18:45, spike jones via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
>
...
>
>>... But I digress.  I am fooling with a Kalman filter which puts our 
> available observables (Atlantic surface temperatures) into my 
> covariance matrix to predict European droughts.  ...
> My Kalman filter is having a great time.  Or rather it would be 
> enjoying itself, if matrix calculus could have a great time.
>
>>... spike
> _______________________________________________


>...Well, London and SE England just got dumped on overnight with
thunderstorms and lighting and 2 or 3 inches of rain. Not enough to refill
reservoirs, but it did cause some localised flooding. Quite annoying if you
are in a drought restrictions area and the streets are flooded!
The UK is quite a small island, but weather conditions vary a lot between
north and south and east and west.
>...Drought areas and flood areas are often quite small (by USA standards).
California is almost twice the size of the UK.

BillK
_______________________________________________


OK, best wishes with that.  You personally are not in danger of an
involuntary swim, ja?

If we didn't know how it works, a Kalman filter could be considered AI: it
learns over time, and gets better at predicting an outcome.  Over time, we
accumulate more data (such as Atlantic surface temperature and rainfall in
England) so the uncertainty in the correlation coefficients in the
covariance matrix improves.  As that covariance matrix gets more and more
accurate, so do the predictions.  In that sense, the Kalman filter learns
from its own past experience and accumulates knowledge of a system it is
trying to predict.

Conclusion: if we didn't know how it works, a Kalman filter would be AI.
Since we know how it "learns" then it is demoted to an algorithm and isn't
actually intelligent.  The term artificial intelligence requires that we
actual intelligent beings don't understand how it works.  Humans somehow
manage to demote software we understand to non-intelligence without
simultaneously demoting ourselves to non-intelligence.  This alone is
sufficient evidence of our actual intelligence.

spike