What's new

Pakistan Defence

VCheng
VCheng
What part of it? Some of my work involves neuroimaging, but that is more of the applied science aspect of it.
_NOBODY_
_NOBODY_
@VCheng I meant in general as I have a question regarding neuroscience.
VCheng
VCheng
What is the question?
_NOBODY_
_NOBODY_
@VCheng From what I understand, the neocortex stores information hierarchically. If I am correct, then what is this hierarchy and how does storing information like this affect the process of processing the outputs relayed by the receptive fields?

1714589134749.jpeg
_NOBODY_
_NOBODY_
Does the hierarchical storage of information go from analysing each individual color at a very tiny singular point to combining these points to perceive an object eventually?
_NOBODY_
_NOBODY_
What is the lowest detail in this supposed hierarchy?
_NOBODY_
_NOBODY_
It's from the book AI 2041: Ten Visions for Our Future.

1714589506790.png
VCheng
VCheng
Sorry, I do not check PDF and related fora very often anymore, hence the delayed reply (emailing me would perhaps be quicker for me to respond). Your question is related to neural network modeling of what our understanding of signal processing is in various areas of the brain, and not really neuroscience, albeit in understanding signal processing by the cortex.
VCheng
VCheng
Much of visual perception, including color, begins in the retina itself, then to the visual cortex and then finally into the higher areas, in a very complex manner. What the brain actually perceives in the end depends on may other related factors, such as the design and orientation of the signal itself, to the emotions it can generate.
VCheng
VCheng
That would be an almost impossibly complex task to model, but a simpler goal would be to limit the modeling to a particular set of constraints based on what the neural network is going to be trained on, for example in a particular area in radiology, looking at mammograms or pulmonary nodules, for instance. Or perhaps you are trying to work out object perception as part of machine vision, for a self-guided robot, say.
VCheng
VCheng
I hope I am making some sense to you, given that I had to make multiple comments due to the 400 character limit.
_NOBODY_
_NOBODY_
@VCheng From what I understand, the first detail that is picked by the receptive fields is the color at a singular point then it goes on from there towards other points which may eventually result in the formation of a basic object like a line or another basic shape.
_NOBODY_
_NOBODY_
Then from there the receptive fields perceive more complex shapes and this process keeps expanding. Am I correct?

Back
Top Bottom