ontolog-forum
[Top] [All Lists]
<prev [Date] next>
[Advanced]
<prev [Thread] next>

Re: [ontolog-forum] Child architecture

from [John Bottoms] [Permanent Link][Original]
To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John Bottoms <john@xxxxxxxxxxxxxxxxxxxx>
Date: 2014年12月27日 12:07:04 -0500
Message-id: <549EE738.4050707@xxxxxxxxxxxxxxxxxxxx>
Philip,

Thank you for your observation, I too had some difficulty understanding Steven's approach. Maybe I still don't understand it fully but let me offer this...

* * * * * * * *
Steven,

Modeling of human perception has been an area of extensive investigation since before I became aware of it in '72. It was in early use in photo work back to the earliest airborne recon work. A few years ago I worked on a system for TV signals that had the goal of determining the type of signal without fully modeling the human visual system. The goal of modeling the human visual system is one of the "holy grail" pursuits that is unlikely to ever be solved because of the complexity of the problem space. People have different visual responses, populations have different names for visual components, noise affects vision differently in different people and the eyes are basically components extensions of the brain and subject to training.

So, what can one do? There was a concerted industry effort in the U.S. and other countries to determine the content and correlation of TV signals against known signals. Why do this:

 º  If you are a cable TV provider, you would like to know if the signal the subscriber sees is close to what is being sent. For example, the quality of a 1950's movie is not the same as that of a more recent digital movie, so knowing the degradation may show that it is not be due to the transmission channel.

 º If you would like to know if someone is pirating a film, then you can compare what is being shared over the Internet with a list of the movies for which you have copyrights. This is the technology that wound up being used in the Six Strikes monitoring of Internet channel content.

If this could be done without modeling the human visual system it could be faster and less expensive. That is what transpired, developed by the various international corporations with assistance from the U.S. Commerce Dept in Boulder using the Commerce Department patent..

The patent is online if you would like to see how they dealt with this problem. I believe future systems will take this approach to perceptual modeling, if the resultant model can meet their requirements.
      http://www.freepatentsonline.com/20070088516.pdf

-John Bottoms
 FirstStar Systems
 Concord, MA USA

On 12/27/2014 8:56 AM, Philip Jackson wrote:
[
To: Steven Ericsson-Zenith, Ph.D.
 
Dear Steven,
 
After reviewing the information you sent, materials at your website, and your correspondence below with Michael Brunnbauer, it is clear you would consider any discussion of biophysical consciousness (or of biophysical sensation, perception, or experience) based on Turing computation to be inadequate.
 
You are developing a continuous mathematical formalism to explain biophysical sensations and responses, and believe the biophysics cannot be Turing computable, at least with adequate efficiency to achieve biophysical power constraints. (Viz. SEZ-2013,p.7, and remarks at end of SEZ-2013 lecture.)
 
To contrast, my thesis discusses an approach toward human-level artificial intelligence which includes artificial consciousness and takes Turing computation as a starting point. Yet the TalaMind approach does not preclude use of continuous computation transcending Turing machines, per section 4.1.2.4. While I think much can be accomplished with Turing computation, I'm open to the possibility that quantum computation and/or continuous computation may be needed to achieve human-level AI.
 
Only a first chapter of your future work entitled 'On the Origin of Experience' has been released to the public. When your work is more fully published, questions and criticisms may be expected from the biophysics and theoretical physics communities, in their forums for discussion. However, it seems clear that discussion of your theories would be out of scope in this forum.
 
It appears your approach will be controversial, at least in the theoretical physics community, given your claim that a valid reformulation of general relativity theory can be developed in which light does not move. (SEZ-2013,p.17) Your approach seems very ambitious, since it appears you aim to derive quantum mechanics and the Standard Model of subatomic particle theory from Einstein's general relativity (SEZ-2013,p.19) and you aim to develop a formalism for biophysical consciousness by adapting Einstein's equations for generality relativity, treating consciousness as a universal primitive similar to gravitation.(Viz. slides 53-59, SEZ-2008; the term 'consciousness' does not appear in SEZ-2013, which refers to 'sense' and 'experience' instead.)
 
So, at this point I am doubtful about the theoretical prospects for your research approach. Still, I wish you all the best in your research.
 
However, further discussion between us is moot at this point. It is clear we need to agree to disagree in this forum about AI and consciousness.
 
Sincerely,
 
Philip C. Jackson, Jr., Ph.D.
 
References:
 
SEZ-2013: Ericsson-Zenith, Steven (2013) On The Origin of Experience - The Shaping Of Sense And The Complex World. Preview, read at Stanford University, November 13th, 2013. https://www.createspace.com/Preview/1137409 and http://youtu.be/zF5Bp_YsZ3M
 
SEZ-2008: Ericsson-Zenith, Steven (2008) A New Kind of Positivism. A Stanford University Seminar, March 13, 2008. http://www.iase.info/2008/03/a-new-kind-of-positivism.html
 
Jackson, P. C. Toward Human-Level Artificial Intelligence – Representation and Computation of Meaning in Natural Language. Doctoral Thesis, Tilburg University, 2014.
http://www.philjackson.prohosting.com/PCJacksonPhDThesisInformation.html
 
 
Date: 2014年12月25日 17:13:52 -0800
From: steven@xxxxxxx
To: ontolog-forum@xxxxxxxxxxxxxxxx
Subject: Re: [ontolog-forum] Child architecture

Should you watch the video? I'd say that it certainly can't hurt.

I like Aaronson's work and find much to agree with. However, It should be clear that I believe, speaking now as a computational theorist with much experience in dealing with parallelism and locality, that we need new ideas. I expect most experienced people in the field will agree with this. The hump that I overcame a decade ago is to accept that there is something new in biophysics and that this is now accessible due to biophysical research funded for other means. IOW, we need new physics and we are informed by biophysical research, primarily funded for medicine. In this Aaronson probably does not agree with me, but you'd have to ask him. 

What do I mean by "non-locality that is evident in biophysics"? I would ask you to sit at a desk with a couple of thumb tacks, one in each hand, and play with them awhile and to tell me if the answer is not readily apparent, or better and more fun, to ask you to listen to Beethoven, view magazines while you masturbate, and eat honey :-)  If the non-locality in biology is not then readily apparent to you then there may indeed be (philosophical) Zombies among us.

The important point is to look at the consequences of sensation, feelings, upon behavior. There is a role for sense in physics, a role that it plays in behavioral outcomes. In my work this is a claim for the primacy of structure, in particular flexible closed structure - that covers every form of biology from bacteria to you and I. It is time for us to put aside the inactive and passive view of what philosophers call "qualia."

With a mathematical characterization of this effect in flexible closed structure - I propose a couple of routes to this, in a state of flux - we can specify a new model of computation able to apply the non-locality involved. Very generally, consider the following a preliminary account of the "symbolic processing" involved. Imagine a holomorphic function that describes a dynamic "shape" upon the surface of such structure (producing an apprehension) and a holomorphic function upon the "opposite" surface that describes the shape of a response and combine the two in a hyper-function. Imagine a simplified bacteria with receptor formation potentials at one end and motor function potentials at the other. My claim is that this mechanism is the basis of all thinking, feeling, and all "life", at all levels.

It also provides a way into new any-scale computational mechanisms in which "recognition," "memory" and "large-scale decision making" comes, essentially, for free in energy terms. 

There are obviously a lot of questions that remain to be answered and a lot more work to do.

I hope this clarifies. 

Regards,
Steven



 

On Wed, Dec 24, 2014 at 12:50 PM, Michael Brunnbauer <brunni@xxxxxxxxxxxx> wrote:

Hello Steven,

Merry Christmas!

Your comments are not very enlightening for me. Should I watch the video?

I cannot possibly comment your views on general relativity but if you do not
aim for non-computable functions but instead for philosophical implications of
computational complexity, you probably have read Scott Aaronson?

 http://arxiv.org/pdf/1108.1791v3.pdf

I'll have to read this again to catch on but my guess is that you disagree
with Aaronson.

What do you mean with "non-locality that is evident in biophysics"? Most
physicists would sell their grandmother for locality.

Do you claim that consciousness harnesses a computational power of the
universe not available to our machines?

Regards,

Michael Brunnbauer

On Wed, Dec 24, 2014 at 10:39:04AM -0800, Steven Ericsson-Zenith wrote:
> Yes indeed, falsification is a part of physics, I argue, although this
> point of view is not very popular among physicists these days.
>
> Until 2010 all of my work was in public. Unfortunately, I discovered that
> this is not sustainable for advanced basic research, early and incomplete
> ideas are too often subject to unwarranted criticism.
>
> My claim is only that the current models of computation are inadequate - in
> this I agree with Penrose. In particular, the solutions to General
> Recognition and decision making across large scale structures are
> impossible to consider.
>
> I make an informal argument, based upon a power analysis for conventional
> computation, that says modern computation is excluded from biophysics.
> There is no "load/store architecture," for example, in biophysics and, if
> you watched my video, at the end I note that memory is "free" in energy
> terms.
>
> So we need something new. That something is based upon a non-locality that
> is evident in biophysics. I hope to illustrate the new physics, along with
> new foundations for logic and apprehension, required to enable new models
> of computation in the book. I do not believe that anything is
> "non-computable," as Penrose does, in this sense.
>
> I should add that the physics I work with is experimental. It is my goal to
> provide viable alternative explanations for the current wealth of 20th
> century physical data, but my focus is the biophysics.
>
> My claim that light is static and the foundation of the gravitational
> effect, appears necessarily controversial, I know. It is, however, based
> upon existing equations in physics. Indeed, it exists in contemporary
> literature as the "heat death" of the universe. I simply use it as a purely
> mathematical starting point from which to reason. If you consider a
> universe consisting only of "photons," in conventional terms "traveling at
> the speed of light" then there is no "time" or "motion."
>
> Why get involved in the physics at this level? Well, it is important that I
> illustrate how we can mathematically unify a viable theory of biophysics
> with the remainder of physics.
>
> The book will be finished and generally available in the middle of the
> coming year. My health allowing.
>
> Regards,
> Steven
>
>
>
>
>
> On Wed, Dec 24, 2014 at 2:42 AM, Michael Brunnbauer <brunni@xxxxxxxxxxxx>
> wrote:
>
> >
> > Hello Steven,
> >
> > On Sun, Dec 21, 2014 at 02:36:48PM -0800, Steven Ericsson-Zenith wrote:
> > > I suggest you familiarize yourself with my main arguments that can be
> > found
> > > in a recent Stanford lecture, available on YouTube, and in written form
> > at
> > > https://www.createspace.com/Preview/1137409???
> >
> > Your arguments are not more clear to me after reading this.
> >
> > Why is the full book only available to "Qualified subscribers"?
> >
> > Do you claim that physics involves a non-computable function?
> >
> > Does your theory make any falsifiable predictions?
> >
> > Regards,
> >
> > Michael Brunnbauer
> >
> > --
> > ++  Michael Brunnbauer
> > ++  netEstate GmbH
> > ++  Geisenhausener Straße 11a
> > ++  81379 München
> > ++  Tel +49 89 32 19 77 80
> > ++  Fax +49 89 32 19 77 89
> > ++  E-Mail brunni@xxxxxxxxxxxx
> > ++  http://www.netestate.de/
> > ++
> > ++  Sitz: München, HRB Nr.142452 (Handelsregister B München)
> > ++  USt-IdNr. DE221033342
> > ++  Geschäftsführer: Michael Brunnbauer, Franz Brunnbauer
> > ++  Prokurist: Dipl. Kfm. (Univ.) Markus Hendel
> >
> >
> > _________________________________________________________________
> > 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
> >
> >

>
> _________________________________________________________________
> 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
>


--
++  Michael Brunnbauer
++  netEstate GmbH
++  Geisenhausener Straße 11a
++  81379 München
++  Tel +49 89 32 19 77 80
++  Fax +49 89 32 19 77 89
++  E-Mail brunni@xxxxxxxxxxxx
++  http://www.netestate.de/
++
++  Sitz: München, HRB Nr.142452 (Handelsregister B München)
++  USt-IdNr. DE221033342
++  Geschäftsführer: Michael Brunnbauer, Franz Brunnbauer
++  Prokurist: Dipl. Kfm. (Univ.) Markus Hendel




_________________________________________________________________
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  (01)
[More with this subject...]
<Prev in Thread] Current Thread [Next in Thread>
Previous by Date: Re: [ontolog-forum] Child architecture , Michael Brunnbauer
Next by Date: Re: [ontolog-forum] Data Silos , Frank Guerino
Previous by Thread: Re: [ontolog-forum] Child architecture , Philip Jackson
Next by Thread: Re: [ontolog-forum] Child architecture , Michael Brunnbauer
Indexes: [Date] [Thread] [Top] [All Lists]

AltStyle によって変換されたページ (->オリジナル) /