Tomography of Consciousness
Christopher McKinstry
Tomography of Consciousness
It took 4 billion years and a great many improbable events for Leonardo da Vinci’s Mona Lisa to appear on the
Earth. It is unlikely that this event could ever again occur independently anywhere again. However, a skilled
artist could copy the Mona Lisa in a few months, a camera in a fraction of a second, and a computer could make
millions of copies in moments. Once something exists (which is always the hardest part), only two things are
required to duplicate it; A medium of sufficient complexity to capture the essential complexity of the
original, and a feedback mechanism that minimizes the differences between the copy being made and the
original. (It is conceded that it would be very difficult to make an ‘exact’ molecular duplicate of the Mona
Lisa, however in this case the essential complexity to be copied is the two dimensional visual image made on
the human retina by the painting, the paintings image. Not it's molecular construction.)
***
It was not long ago that the process of non-destructively imaging the inside of the human body was thought to
be impossible. Now millions of such images are created every year using a host of techniques including CAT,
MRI, PET, SPECT, etc. The tomographic process with which all these types of images are created provides us
with a model for imaging and duplicating the common awareness of self and environment we like to call human
consciousness.
In creating a CAT scan image, many x-ray samples are taken from many different vantage points. The resultant
data is then statistically correlated, essentially extracting the common pattern that could have ‘caused’ each
individual sample, without contradicting any other individual sample. For example, one sample yielding an
x-ray intensity of .33 (on a scale 0.0 to 1.0) would by itself indicate it was ‘caused’ by an object with a
uniform x-ray density of .33. As more and more samples are taken from more and more view points the overall
quality of the entire image is slightly improved and the .33 figure is revised to remain consistent with all
subsequent samples. Millions of these samples yield the near perfect tomographic images of the internal
structure of people and objects so common today.
***
Imagine a simple database with millions of ‘stimulus/response’ pairs; where each stimulus would be a statement
of consensus fact (facts that most people agree on, most of the time), such as ‘The day time sky is usually
blue.’, and where each response could only be ‘true’ or ‘false’. Each of these pairs is equivalent tomographic
sample.
As with a CAT scan, one sample will not tell us anything of use, nor will ten, nor will ten thousand. Millions
however will yield a high resolution binary tomographic ‘image’ of the common cause of all the individual
samples; an image of human consciousness which could be used to project a true artificial intelligence into a
computer.
An artificial intelligence constructed bottom-up from a binary tomographic image feedback process would be
indistinguishable from a human being when asked questions that can be answered in a binary fashion. Such an
entity could communicate awareness of self and environment and would thus qualify as being conscious. Consider
the following brief dialogue between such a hypothetical system and a person:
Are you alive?
yes
Do you think?
yes
Have you ever felt lonely?
yes
Have you ever not felt lonely?
yes
Do you like to feel pain?
no
does 2 + 2 = 4 on mars?
yes
Do people have three arms?
no
Is it possible to be in two places at once?
no
With such a system, a person could ask it questions for a lifetime and not find anything inconsistent with
interaction with a living human being. When a question is asked that is ambiguous or not answerable in a
yes/no fashion, it does not matter what the system responds so long as it is consistent with how it has
responded to similar questions in the past (which is the same expectation we would have of any human). Such a
system could not be distinguished from a human being locked in a box, who is only permitted a binary response
to any question.
***
There exists a project called MISTIC (Minimum Intelligent Signal Test Item Corpus) charged with collecting the
necessary stimulus/response pairs to create a database that can be used as a feedback function for training
systems (such as neural nets, or other statistical correlation system) to act/become conscious.
Please contribute as many stimulus/response pairs as you can think of (via email to ch...@clickable.com) and
soon using the form that will be accessable at: http://www.clickable.com/mist.html
Millions are needed.
thanks for your time.
chris.
--
K. Christopher McKinstry
-All science is the study of perception.
-AI researchers, Win $10,000 http://www.clickable.com/mist.html