Newsgroups: alt.ufo.reports
From: kymhorsell@gmail.com
Subject: predicting the front of the elephant

[uploaded 30 times; last 26/10/2024]

I sometimes get a bit blaze about what the AI's I tinker with can do. But
it's really pretty remarkable. To a non stats-head it either means nothing
or is akin to black magic.

AI programs can be frustrating to work with. They can figure out a lot
but unless the design is very very careful they will not be able to
explain a lot in terms any human can understand. I can imagine that a
civilisation could easily dead-end itself relying too much on bad AI software.
Not that that could every happen to us! ;)

For some months I've been tinkering with new techniques to predict the
ebb and flow of certain activity. A couple years ago when I started out
looking at this stuff I thought you might be lucky to get within an
order of magnitude in predicting how many reports might come in next
month based on whatever data you could dig up that seemed related.
But the latest programs far exceed those old expectations.

The latest round of models the s/w has cooked up are based on
"planetary" movements. It turns out the month-to-month activity can be
predicted very closely over at least the next days by the
movements of just a few key objects in our solar system.

The best result in the latest batch -- that tries to predict in
advance what coming month-by-month activity will be -- is here:

<kym.massbus.org/UFO/UFO-WEATHER/save2/>.

As usual, it's like an intelligence test to dig your way through the
results. But we all like a knotty after-dinner puzzle, don't we?

The model uses the "phase" of 8 asteroids/comets from data uploaded
from the good old Horizons system at JPL. The phase is roughly the
angle as seen from some object in the solar system between the Earth
and the Sun. Back here on Earth that relates to how much of the
object looks lit by the sun as viewed from the ground here.

But it turns out to have other uses. E.g. in planning an
interplanetary trip between there and here.

The model, above, is based on a variation of the phase angle for the 8
objects in the model -- the "Phase Angle Bisector" (PAB). Head
spinning yet? Wait. The phase is the angle as seen from some other
planet between the Earth and Sun at a particular date. The *bisector*
is the direction exactly between the Earth and the Sun as viewed by
that other planet. Roughly, very roughly, it's the direction you have
to go to get from there to here in a vaguely reasonable time.

Not at all funny that these PAB numbers predict UFO sightings to a
high degree.

The list of objects in the model(s) is not particularly important. It
turns out you can take almost any set of the known ~2 million objects
being tracked and get the same kind of model. The one listed above is
just the best one found in the latest twiddling.

It's hard to describe how well the model works. To number-heads we
look at the output from the program. For normal people they might
relate to the graph of the predictions and original data a bit better.
That's here:

<kym.massbus.org/UFO/UFO-WEATHER/save2/oe.gif>.

The plot shows in green the monthly "activity" numbers between 2006
and 2023. The data comes from the NUFORC. The s/w massages the
numbers a bit to make the model-fitting a bit more precise. We know
the number of sightings in a month is only a rough estimate of
"activity". Someone may forget exactly when they saw something and
just guess at the date. Some of these reports come in months or years
after the events. Something about stigma.

So the numbers represented by the green bars are smoothed estimates of
the annual rate of activity based on a given month's number of
sightings reports as at the start of 2023.

The dark blue line in the plot is the model estimate. You can see the model
is a bit of a wavy line. Since the PAB numbers from orbiting planets
go up and down, adding a whole bunch of them together will have this
appearance. Which is great because we can see activity goes up and
down in cycles, too.

Now here comes the knock-your-socks-off moment.

Because we have 2 million objects to choose from you might think it
was a slam-dunk that you could find a bunch of them that closely
parallels anything at all, even random sightings reports sent in by a
bunch of pilots and ordinary citizens looking up from their back
porches on a weekend evening.

And that is perfectly true.

But data science takes that into account by "information hiding". We
give the stats program only PART of the data. It figures out the
pattern from e.g. 1/2 or 1/3 of the available data. Then we TEST the
model the program has found by seeing how well it matched the 50-67%
of the data we DID NOT GIVE IT.

It's like showing the program the back of an elephant and seeing
whether it can figure out the front of the elephant must have a trunk.

The spooky part of data science is -- current methods can do it! The plot, 
above, shows a thick blue line on the left 1/2 along
the bottom of the X axis. This is the part of the data the program was
given to predict the other 1/2 of the data. And the match is near perfect.

From a few wavy lines on the left 1/2 (the "bum" end of the elephant) it 
predicted the "pandemic peak" on the right half. There's your trunk!

This is the difference between "curve fitting" -- a thing still
beloved of traditional scientists and wanna-be scientists -- and
"machine learning". In curve fitting the model you come up with has
only remembered the numbers you gave it. It has not "generalised" the
data to figure out what could be true in slightly different
circumstances. In ML the model has "digested" the data it was given
and found out the rules used to generate it, and can then go off to
slightly or even moderately different circumstances and make a
reasonable prediction about what will happen.

From the bum-end of an
elephant it can draw a picture of the trunk without having see one before.

Spooky stuff. But just a part of what AI can do. And probably only the
tip of the iceberg when it comes to what Our Friends From Frolix 9 can do.
Predict tomorrow's lottery numbers? Not a problem. We are already building
machines that can do that, if you just ask them nice. :)

--
"Nothing in life is to be feared, it is only to be understood.
Now is the time to understand more, so that we may fear less."
- Marie Curie

The most extreme life-forms in the universe
New Scientist, 26 June 2008
There's hardly a niche on Earth that hasn't been colonised. Life can be
found in scalding, acidic hot pools, in the driest deserts, and in ...
[Interestingly, if life is *not* found in the warm salty sub-surface
oceans of some of our system's moons it gives more weight to the idea
that life could not have formed spontaneously on Earth but came from
"outside" e.g. via meteorites aka Panspermia].

[Priorities!]
Pentagon's UFO office had only THREE staffers until recent weeks
Daily Mail, 27 Feb 2023 02:08Z
The All-domain Anomaly Resolution Office (AARO) was set up by the Pentagon
in July 2022 to probe UFO sightings in the US.

California Glider Pilot Caught Possible UFO Footage Mid-Flight
Exemplore on MSN, 27 Feb 2023 13:07Z
On the flip side are videos such as this one, where an innocent glider pilot
sees something strange outside his window and ...

Pentagon Hearing Air Force Veterans With UFO Stories
Newsmax, 22 Feb 2023 0:17Z
Two Air Force veterans told the Daily Mail they testified this month
before a Department of Defense office allegedly looking into reports
of unidentified...
[Among other things some testified about an incident where missiles in
a silo were "turned off" during an attempted visit by some strange
bright red light hovering nr the main gate].

As the UFO mystery deepens, scientists must engage with the data
The Hill, 21 Feb 2023 17:39Z
When will scientists reject the government-fueled stigma that long hindered
sober, objective study of UAP and grapple with ...

Leading astronomer tells of 40 UFO sightings in Roscommon - and mysterious
objects come "in waves"
Irish Daily Mirror, 20 Feb 2023 20:18Z
[Dr Eamonn Ansbro] has revealed that he has recorded around 40 UFO
sightings at his advanced space observatory in ...

Mystery airships that filled skies of US still unexplained more than 100
years on
Daily Express, 19 Feb 2023 06:44Z
It was a bright light, too big for a balloon' - this is how one newspaper
described a UFO sighting, one of the dozens that ...

US ends search for two UFOs shot down over Alaska and Canada: report
New York Post, 18 Feb 2023 04:38Z
The search for two mysterious objects shot down by the US military has
concluded with no debris being found, according to a report.
