I wonder why we presume that a "theory of everything" has to be grounded in
physics. What about the approach that every "thing" is actually only knowable
as a symbolic representation as some kind of concept -- including fundamental
particles? Seen this way -- a "theory of everything" involves a "universal
theory of concepts" and the issue becomes reliably connecting the abstract
symbolic representation ("the concept") to the "actual empirical object" --
like a fundamental particle -- or any other discernible and bounded/distinct
object. (01)
I like physicist Arthur Eddington's suggestion in his "Fundamental Theory"
(1946): "All the laws of nature that are usually classed as fundamental can be
foreseen wholly from epistemological considerations." (02)
My own feeling is -- what is needed today is a new foundation for mathematics
itself -- a new ontology of mathematical fundamentals -- that interpret in
mathematical terms what Eddington was calling "epistemological" (03)
I like an approach that says something like "all concepts can be constructed
out of dimensions and the fundamental constructive element is a cut (as per
Dedekind) or distinction." You can build any concept taking that approach and
it is 100% linearly recursive and extremely simple. (04)
For me, all of this floats in a framework defined by John Sowa's basic
proposition, as I read it: "concepts are discrete, reality is continuous". We
need to generalize the framework that contains this fundamental truth, and show
how all mathematics -- arithmetic, real numbers, continuous variation --
emerges as attributes of this containing framework. Distinctions within this
framework become "all the concepts in reality". (05)
Do this in a top-down way -- and you get something like the model described in
the Numenta white paper on hierarchical memory, which mostly describes simple
linear recursion. Build it up from the bottom (based on observable empiricism)
-- and you get something like John Sowa's "semantic networks". (06)
I don't know if "neurons" are organized in the way Numenta suggests -- but I'd
say concepts can be -- and neurons support conceptual processing. (07)
We should connect these approaches. We need a "transcendental container" that
can hold all of this in one framework. This mathematical object might be
fairly simple -- if dimensionally a little tricky -- and emerge as the
conceptual foundations for a new and very-much simplified and integrated way to
understand reality. (08)
- Bruce (09)
PS -- ideas rule the world. We need to understand ideas -- good ones, bad
ones, confused ones, crazy ones. It's ideas that cause teenage girls to join
ISIS. (010)
-----Original Message-----
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John F Sowa
Sent: Tuesday, March 03, 2015 5:12 AM
To: [ontolog-forum]
Subject: [ontolog-forum] Grand Unified Theories (011)
Physicists are the closest to finding a Grand Unified Theory (GUT) of
everything. But every time they find one, it opens up far more mysteries than
it solves. (012)
Meanwhile, the very many practical applications use old theories that are known
to be be inadequate in the details. Good old-fashioned Newtonian mechanics is
a prime example. For big things moving at normal speeds, GONM is the first
choice. (013)
Two professors at NYU -- the psycholinguist Gary Marcus and the AI expert
Earnest Davis -- wrote a review of attempts to find a GUT about intelligence,
human or machine:
http://www.newyorker.com/tech/elements/a-grand-unified-theory-of-everything (014)
Opening paragraph:
> Here’s a way to make a lot of money. Publish a speculative scientific
> article with equations nobody understands, put out a press release,
> throw in a few credentials (say, a degree from Harvard or MIT), and
> get a few bloggers to spread the word. In the meantime, quietly start
> a company based on the idea—the grander, the better. (015)
Marcus also wrote an article "Steamrolled by Big Data", which discusses a GUT
by Jeff Hawkins, which is supposed to do everything:
http://www.newyorker.com/tech/elements/steamrolled-by-big-data (016)
Quotation:
> As one skeptical industry insider, Anthony Nyström, of the Web
> software company Intridea, put it to me, selling Big Data is a great
> gig for charlatans, because they never have to admit to being wrong.
> “If their system fails to provide predictive insight, it’s not their
> models, it’s an issue with your data.” You didn’t have enough data,
> there was too much noise, you measured the wrong things. The list of
> excuses can be long. (017)
I went to the web site for Hawkins' company, Numenta, where I found the
following whitepaper about their Hierarchical Temporal Memory:
http://numenta.com/assets/pdf/whitepapers/hierarchical-temporal-memory-cortical-learning-algorithm-0.2.1-en.pdf (018)
This draft is version 0.2.1 from Sept. 2011. All the newer things are videos,
mostly by Jeff H., who talks very fast. I also found an MS thesis from 2011 by
Ryan Price. (019)
If anybody can find anything newer and better, please let me know. (020)
John (021)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/ Community Wiki:
http://ontolog.cim3.net/wiki/ To join:
http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (022)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (023)
|