Mike Tintner wrote:
Richard,
Thanks for response. But it surely *is* still a puzzle as to how and
indeed where that distorted image on the retina gets rectified and
raises major questions about vision. No one as, I understand, has the
answer. I am too ignorant to have a POV here - but my general
experience is that people/scientists faced by major unsolved problems,
tend to proceed as if they do not exist (or will be solved in good
time), when they should be racked by them.
Not true. The image from the retina is not distorted in the visual
areas. Evolution is imperfect. The visual cortex is complex and over
convoluted that simpler methods would do better. But this does not imply
that we should avoid spatial processing and vision. We need this
research to theorize cognition, in general.
I'm starting to get an unified model of our brain.
Appraisal systems are required for cognition. How do you deduce that
"taxation is theft" when you know the meanings of taxation and theft?
You cannot deduce this statement without emotion. Because "theft" is
used in negative connotations, it is reasonable to write "taxation is
theft".
When you perform mathematics, you perform unconscious, intuitive
reasoning. When you do repetitive tasks, memory tends to get
consolidated and "optimized" to the basal ganglia. (Anderson, 1982) This
is why Parkinson's patients are poor at applying rules, language and
goal oriented tasks, while Huntington's patents tend to perform better
at these tasks because they have better implicit memory. (Lieberman, 2000)
The basal ganglia has a major effect on intuition (Lieberman, 2000). The
basal ganglia may perform automatic linking of attributes. (Anderson,
1974; Cheng, Saleem, & Tanaka, 1997) The basal ganglia may let the
person choose a strategy for learning.
Hofstadter had it wrong. Abstration, generally, is an explicit,
goal-oriented association search. Rule-based matching is goal-oriented.
(Smith et al., 1998) This is supported by the studies that suggested
that the PFC is important for category learning (Milner, 1963), PFC,
left frontal and temporal regions for semantic processing (Martin &
Chao, 2001; Gabrieli, Poldrack, & Desmond, 1998) and impaired
categorization in the frontal patients (Ashby, 2005). The lateral
temporal cortex and the medial prefrontal cortex are coactivated when
semantic memory is used (Lieberman, 2000; Lee, Robbins, Graham, & Owen,
2002; Xu et al., 2002).
This PFC may function as a storage for temporal information, sensory
integration and context dependent recall.
There is no general abstraction mechanism, except a few simple
dirty-lookups.
REM sleep may perform implicit memory consolidation. Studies shown that
the ventral prefrontal cortex performs appraisal, motivation,
stereotypes, semantic processing, etc.
Humans cannot even solve the SLAM problem without measuring. Humans
learn 3D vision using associations with the help of place cells in the
hippocampi. People who are not exposed to buildings are much less likely
to see the Muller-Lyer illusion. (Segall, et al., 1963) When humans see
a picture, they subconsciously convert the 2D image to three-dimensional
using neural associations, and then converting the image to place cells
and spatial cells in the hippocampi. Visual processing, such as 3D
visual reconstruction, requires the inferior temporal cortex.
(Lieberman, 2002)
Women tend to have more white matter, which may explain their advantage
to verbal processing compared to men. This is also confirmed by Doreen
Kimura, which she concluded that women have faster array-recall,
categorization, linguistic constraint satisfaction, and visual searching
tasks.
The lateral temporal lobe, and the temporal lobe, in general, performs
semantic processing (Garrard & Hodges, 2000; Mummery et al., 2000),
priming (Rolls, 1999; Tomita et al., 1999), and pattern matching
(Liberman et al., 2002; Evans, 1989). Furthermore, visual matching is
commonly known to involve the ventral stream. These parts of the brain
may be highly associational. The inferior temporal lcortex performs
priming (Tyler & Moss, 1998).
We are more likely to recall events that occurred temporally close.
That's why we recall more things in STM than others. Implicit memory is
equivalent to long term memory: you cannot access implicit memory
directly, and you cannot also access long term memory directly.
Negative priming may be a symptom of neural adaptation. The filling-in
effect, is not a visual phenomenon (Komatsu, 2004), therefore may be an
example of the effects of neural adaptation. Negative priming evolved
probably because it ignores distractive repetitive stimuli, so you can
concentrate on more important tasks.
Gestalt perception, requires semantic processing. (Eherenstein, 2003)
Episodic memory has an important effect in humans. This kind of memory
may be formed by temporal associations.
Priming is "warming up" associated neurons so you can recall more
accurately and produce more accurate context dependent pattern matching.
When automatically match primed associations. When you learn something
temporally together, you seem to recall them together more. This is why
"recalling digits from one through ten" is easier than "recalling jobs
that begins with the letter 'A'".
When you recall "jobs that starts with the letter 'A'", you don't
automatically recall through an index. You recall by (unconsciously)
generating unrelated associations by priming, like airplane -> air
piolot -> astronaut -> science-fiction movies -> actor -> artist. This
is a criticism of ontological knowledge bases because ontological
knowledge bases must store the letter 'A' in each job that begins with
the letter 'A' before searching. Humans do not do that because it takes
up too much memory.
"Artificial general intelligence (AGI) has been solved by us using a
transactional database. We thus conclude that we cannot be overridden by
the smartness of the problem. If we then transform these matters for the
betterness of the human population, we would be better off." This is a
sentence that I randomly generated. Notice how "transaction,"
"overridden" and "transform" are similar even though they are randomly
generated.
When you recall "jobs that starts with the letter 'A'", you don't use
any index. You go through a series of random associations, trying to
find them in your brain. This is a time consuming process.
So I'm building an AGI based on unconsciously priming (evidence suggest
we have more unconscious) and associationism.
-------------------------------------------
agi
Archives: http://www.listbox.com/member/archive/303/=now
RSS Feed: http://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b
Powered by Listbox: http://www.listbox.com
