6½ Fundamental Principles Of Cognition

[bfrs]

http://www.foundalis.com/res/poc/PrinciplesOfCognition.htm

This is Harry Foundalis' attempt to identify the fundamental
principles of cognition. Harry worked with Douglas Hofstadter and
wrote a program, Phaeaco, which solves Bongard problems (http://
www.foundalis.com/res/diss_research.html). According to Hofstadter the
ability to solve Bongard problems is one milestone that an AGI should
cross.

Please give your comments and critique of these fundamental
principles.

[Ben Goertzel]

All these are clearly important aspects of cognition, but do you have
a clear argument written somewhere regarding why they should be
considered the foundational aspects (instead of just parts of a longer
list?)

-- Ben

> --
> You received this message because you are subscribed to the Google Groups "opencog" group.
> To post to this group, send email to ope...@googlegroups.com.
> To unsubscribe from this group, send email to opencog+u...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/opencog?hl=en.
>

--
Ben Goertzel, PhD
http://goertzel.org

"My humanity is a constant self-overcoming" -- Friedrich Nietzsche

[bfrs]

"All these are clearly important aspects of cognition, but do you have
a clear argument written somewhere regarding why they should be
considered the foundational aspects (instead of just parts of a longer
list?)"

This is something that I would also like to know. I have sent an email
to Harry Foundalis asking him to participate in this discussion and I
hope he joins.

Meanwhile here's my take on it:

Harry says that he was motivated to write his "Fundamental Principles"
after reading Jeff Hawkins' "On Intelligence", which is grounded on
Vernon Mountcastle's hypothesis that the neocortex is running the same
"algorithm" for vision, audio, motor skills, language...everything.
Mountcastle's hypothesis was based on his observations of uniform
cortical anatomy, and today is backed by much evidence, such as the
experiments on ferrets, where the optic nerve was severed and
connected to the aural cortex, the infamous BrainPort tounge vision
device, Henry Markram's blue brain project studying cortical columns,
etc.

Hawkins' book describes what he thinks this uniform algorithm is,
which he later formalized with Dileep George as Hierarchical Temporal
Memory (HTM).

Hofstadter doesn't seem to be aware of Mountcastle's hypothesis
(atleast not when he wrote GEB and some of his other famous books),
but seems to have independently come up with an equivalent notion,
*analogy as the core of cognition*. Harry's essay is his elaboration
on the notion that a few general principles underlie all
intelligences.

Andrew Ng of Stanford has also come around to this idea and his latest
research on deep learning networks is in the same direction as HTMs.
Also it goes without saying that Itamar Arel's (and now opencog's)
DeSTIN are similar approaches.

Reading Hawkins, Hofstadter, Harry, Ng and others*, one sees them
mentioning more or less the same few principles over and over again
reinforcing the idea that a few principles underly all intelligences.

I feel that some of the fundamental principles that Harry mentions can
be subsumed by others, and also that a some things which I consider
fundamental are missing, but more on these later.

---
* Ben, I have still to complete reading your group's CogBot proposal,
where you emphasize the concept of cognitive synergy, which is a
somewhat new perspective to me.