The take away was that our models of ANNs benefit from considering more that just neural signaling (i.e. solution contents and fluid dynamics). Is anyone aware of modern work expanding on that front?
The take away was that our models of ANNs benefit from considering more that just neural signaling (i.e. solution contents and fluid dynamics). Is anyone aware of modern work expanding on that front?
I (no expert in neuroscience) had come across a neuroscience article in around 2004, which I could not find again, and on which I never came across anything else either. This said that myelination happens differently in different parts of the brain, culminating everywhere by adulthood. I had found that very interesting. If someone knows more about this, I would like to hear. :-)
I find the whole xor result to be pretty irrelevant, which I think is representative of the field writ large.
> Is anyone aware of modern work expanding on that front?
I am curious about this as well, as well as the general state of biological inspiration in the field. My impression was that biological analogy had largely been pretty much completely moved away from, except as someplace to take inspiration from in terms of what optimization can achieve.
Basically that Hopfield networks couldn't learn XOR.
The significance is that it moved us away from biologically plausible networks and precipitated back propagation which was the dominant learning paradigm in neural networks and mostly remains so.
[0] https://alan.do/minskys-and-or-theorem-a-single-perceptron-s...
What I demonstrated was up to 6 dimensional XOR functions being learned in Hopfield networks using Hebbian learning with an additional learning term. I ran a "wave of learning" over the network modifying that terms. In a sense it is a layer free network because no explicit later exists but in another sense the modulation provided a transient layers since the learning term caused different parts of the network to learn at different times out of sync with one another.
[0] https://drive.google.com/file/d/1R4mS_4vzf5mycQez76O_mmpQDpT...
To note, I'm really curious how a no-to-low inflammatory vs. high inflammatory diet would impact this visual; likewise seeing differences between short-term and long-term, and things like before and after if switching between/cleaning up a diet or going from low inflammation to high inflammation over 1-5-10 years.