Another Republican debate, Buddhism,
UN deep learning, just daily stuff
My post yesterday on dreams
was an organized presentation of basic material, linking science and experience
and larger issues. Today is basically just a personal diary kind of thing, on
events of the day. (In parallel with this, I have a scientific journal which I
do not post, for multiple reasons).
Themes of the past day or two
– another Republican debate, a discussion of deep learning and political
prediction on the listserv of the Millennium Project, some notes about Buddhism
and the watch, a couple of mundane dreams, a doctor’s visit later today.
Republican debate
---------------------
Maybe the new Republican
debate was a notable step up from the previous ones – even though we almost
couldn’t get to see it! We do not subscribe to Fox business channel, and have
no intention of doing so – and Verizon FiOS welched on its promise to make it
available on a one-time basis for seeing the debates. So we saw it on streaming
– learning how to really use internet to see a TV program, which I haven’t done
so much before.
I now think: “only 5 to 3
against sanity, not so bad.” But the 5 were still deeply disturbing. Some
sounded god in a way, and you could feel the sincerity and resonance, just as
one can feel great sincerity in some asylum inmates. So far out of touch with
reality that it is deeply scary. For example – I did notice some interesting
stuff in the earlier TV interview of “W” about the new book based on his father’s
diaries, but Jeb sounded as if HE didn’t learn any of the lessons the others
did. The book mentioned Cheney building his own “empire” (not something to
underestimate even now!), but Jeb and his sycophants all stressed the
importance of “making it clear that we are the leader” and “pleasing those
wonderful imams who just want to be our friends.” Jeb did so in argument
against Trump, who was one of the three who showed glimmers of sanity. In
effect, those three gave a feeling they had actually set foot on planet earth,
and MIGHT not just blow up the world as soon as their hand would be in the
right place. (Aren’t Jeb and Marco at least stable and reliable?) Not when they
are ready to accommodate folks who want to create a war between US and Israel
versus Russia and Iran.
Side note: I am hoping Obama
and Netanyahu had some interesting side conversations about the Grand Mufti and
the realities of the Third Caliphate movement. Who knows?
The three.
After the debate, I still
feel that I would not be able to be truly open-minded, even if I tried, if the
Republicans nominate anyone but Kasich or Trump. Both had moments of really
being in touch. That doesn’t mean all moments, but enough to engage as Quakers
normally try to. In retrospect, Rand Paul also had a good night. In fact...
Rand Paul gave a great example of how people can learn at times to rise to a
higher plane, and that is what we need most in the world. I still wouldn’t
consider voting for him as President, since I know from previous discussions of
other subjects that he too could blow it all, but if all I knew of was last
night I might have real hope for him (as I did when he stood up for personal
freedom that day on the floor of the Senate).
A grim smile came to my face
when they debated who should get credit for stopping Hilary Clinton’s earlier
health care bill. I know the answer to that more than they did! I know because
I worked for a year in the office of Senator Specter, the guy who did that. His
main conference room displayed the gigantic flow chart which turned him against
Hilary’s proposal, and which he used to get the rest of the Senate against it
as well, and many of the folks there told me about those days. In the end, it
was the Club for Growth (which Huckabee calls “The Club for Greed”) which purged
him, using hardball questionable tactics (like sending violent provocateurs to
his town meetings) to do so. Some folks accused him of being a “rhino,” which
means he thought for himself, and was not a mindless puppet for psychotic
outside forces. He was not the only one purged, nor was I. The empire marches
on. I remember what Icahn said about Trump as his best hope for a “new Teddy
Roosevelt” to get the special interests and evil empires under control. Kasich
was the only one who sounded as if he might have ever paid attention to things
that Yeshua ben Josef (aka Jesus Christ) actually said about some important
things we all should pay attention to. (I don’t go so far as all of the
things... but there are some really important basics.)
************************************
**********************************
Millennium project and deep
learning:
For everyday life, maybe that
would be the subject next most interesting to some of you.
http://www.millennium-project.org/
: The Millennium project, formerly part of the United Nations University but
now a more truly international NGO kind of organization, is best known for putting
out an integrated report every year, which consistently tops the lists of
futurist publications, on what we know now about the future of humanity – drawing
on analysis from “nodes” all over the earth. They recently posted to the
planning committee:
---------------------
During the last meeting of The Millennium Project in San
Francisco this past July, there was some discussion about integrating AI into
MP’s work. I will go to a session on using IBM’s Watson tomorrow. Today Google
announce it is giving access to its deep learning engine for free... This AI
software is called TensorFlow...Google believes it can accelerate the evolution
of AI this way
“What we’re hoping is
that the community adopts this as a good way of expressing machine learning
algorithms of lots of different types, and also contributes to building and
improving [TensorFlow] in lots of different and interesting ways,” says Jeff
Dean, one of Google’s most important engineers and a key player in the rise of
its deep learning tech.
Google built the
underlying TensorFlow with C++ , but coders can also Python. You can get
started using it at:http://www.tensorflow.org/get_
--------------------------
I replied:
This is too close to home to ignore. Just FYI --
-- My younger daughter is part
of that group at google
-- Watson is more like old AI
(expert systems) than neural nets or even
the usual machine
learning. When so many futurists lauded Watson, IBM restructured around
Watson, and, ironically, reassigned its more modern and also its more neural
people to that project, losing much of its advantage. The new IBM game plan
included a couple of physically hard special servers supposed to run the US
government, and some special contacts with NSA, but as issues emerged...
the stock dropped.
-- The "new deep
learning" can certainly be traced back to papers I wrote in 1988 and
earlier. "Deep learning" is often defined "more than one or two
layers in the neural network"; my 1988 paper was the first to show the
general N-layered structure, discussed its advantages and exactly how to
adapt it. ("Generalized MLP.") It also gave the main additional features
needed to get the first wave of new capabilities.
-- Deep learning basically
languished until 2007 -- well documented but used only in specialized projects
-- because of deeply rooted ideologies and vested interests opposed to neural
networks, vested interests still powerful in important places in DC. But in
2007 I sold NSF on a new one-year funding initiative, which I led:
As part
of that I funded LeCun and Ng to do the larger-scale demonstration which LeCun
and I certainly already knew would work, which broke world records in image
recognition, speech recognition in natural language -- which led directly to
the new tide of interests. The NSF abstract is at:
However,
it is a lot of unformed hype or used car selling when people talk about
building systems as intelligent, say, as a mouse brain within the next 20
years. For the latest ground-floor reality, see:
http://arxiv.org/abs/1404.0554
Perhaps
it is just as well that people are far from building true AI, or systems more
intelligent and conscious than human brains, even though we now know the
pathway to do so. Well-crafted lookup systems like google search, Watson, with
pieces upgraded by a proper use of simple neural networks, can still be very
useful, so long as we do not give them too much power or believe them too
much. A nice slide show on the actual history of new products and the
huge risks of real AI and neurotechnology is posted near the top at www.werbos.com/Mind.htm.
-- I am
very sorry I have yet to update www.werbos.com/mind.htm to link to three new papers this year
which take some aspects of this to new levels, one in the book by Freeman and
Kozma on the brain, one in Automatica, and one in press in Quantum Information
Processing.
----------------------------------------
A further reply:
Hi, ... !
I am always happy to hear
from you...
I know of two complementary
technologies which could revolutionize predetection, but lately I feel almost
paralyzed by the issue of trust.
In a way, that is a kind of
extension of the time when I saw Terminator II back in 1991, and realized just
how close I personally was to causing that kind of outcome. (Shadows of that
possible future still show up in our area, much weaker but quite real, from
time to time.) But more and more possibilities for adverse effects and misuse
present themselves to me, quite graphically.
What would it be like to
work for NSA, for example, and one day discover that the guy who was supposed
to be your channel to the President was actually a channel to the guys behind
ISIL? What would you want to send on that channel?
The safer technology is the
less powerful, and probably too futuristic and limited to be of anything
but amusement value to you. Still, since you have enjoyed a bit of time stuff
in the past, I attach the proofs of a paper in press, which cites the work on "ghost
imaging." I would really enjoy it if people go ahead and try out triphoton
ghost imaging on a telescope looking at the sun... and discover after the fact
that they are seeing an image of the sun eight minutes AFTER the time when the
image is recorded. (Flares and sunspots move fast enough that this might
well be possible.) In principle, with slow light and curved paths, it could be
used to image places on earth in the future..
like maybe NORAD's early
warning board.
For nuclear terrorism... I
led the joint NSF-DHS joint initiative on research to better cope with that
threat. It was quite an eye-opener. I knew things were bad, but actually they
were worse. Still, the availability of nuclear material and technology is a
powerful limiter, the most important limiter. Thus improvement of the
nonnuclear, non-CO2 base for generating electricity
ALL OVER THE WORLD really
should get a higher priority than it does now. There is a whole lot more that
we could be doing, that we aren't doing, anywhere on earth. In some ways, China
has actually been the most responsible nation on earth in that area... but the
world needs more than that, and their capabilities are not what ours could be
if only we could rise to the challenge.
Deep cultural issues are
also a crucial driver... but this email is probably too long already.
*************** Buddhism,
Dreams and the Watch
The term “the watch” moved
quickly to prominence in my thinking a week or two ago, when a deeply intuitive
person in the local Meeting and Yeshua both drew my attention to it, and what
it means. Can I live up that mandate? Who knows. It’s pretty scary and
overwhelming, as I may have mentioned above. I don’t have quite the level of
strength to just naturally glide into such a role, and I have a whole lot of skepticism
about absolutely everything – even about my own new thinking about “the law of
everything,” which only sounds conservative if you don’t understand about
emergent phenomena.
And so... Buddhism is BOTH a
subject of the watch, AND a natural reaction to feeling overwhelmed and beset
by “damned if you do, damned if you don’t, how real is ANYTHING?” A few years
ago, I started to notice how respect BOTH for Jesus AND for Buddha had grown
naturally in my mind... not for crazy formal organizations, which sometimes see
as bad as the temple of the green goat in Chengdu which claims to represent
Taoism... but the real people and the real thoughts.
Just a few weeks ago, I was
musing... for almost any subject I think serious about, it’s strange how human
representatives of the subject pop up in my life... not avatars exactly, but a
little like that... Yeshua being an obvious and fascinating example, but there
was a kid who did pretty good Einstein imitation (not so easy to fake higher
dimensional geometry, folks!), and many others... so why no one even remotely
like a real Buddha? (Yes, lots of Zen enthusiasts, and a House reception for
the Dalai Lama, but...). But then I
remember an “assumption” not-dream I had years ago, when a misguided guy was
trying to reach me on the astral plane with bad intent, and was ever so
frustrated he could not find me anywhere... when I was laughing to myself since
I was in fact right inside him at the time, looking through his eyes and seeing
his mind. (Yes, I remember names and the tactics of the conflict and so on.)
So: did I look in the wrong place? Maybe. But I firmly commit myself to caring
about “light, life and love”, and the deeper utility signal which that
characterizes, and Not to any philosophy which gives up. But sometimes it is
really hard to resist thinking and even behaving like a true, core Buddhist. (No,
not unto vegetarianism... but I have resolved that on any day where my weight
exceeds 170 pounds, even fully clothed, I will behave like a Jain, more or
less. No fatty meat, but if Luda offers me salmon roe, green cheese and
porto... I won’t turn her down.)
One day... somewhat
overwhelmed by the watch and by the scary trends of our world... we walked a
couple of miles to the Arlington central library, and I followed my usual
routine for that... first check new books, then up to the magazine section,
where first priority is to Scientific American, Science News (despite my
caveats about what to believe there), and Tricycle, the magazine of the
American Buddhist community.
More later. Off to doctor.
No comments:
Post a Comment