Note: I'm new to neuroscience and AI. I've read few articles about AI and studied them further. So I have a question: Human neurons release NT. The amount of NT released changes the frequency the next neuron is stimulated. AI uses sigmoid neurons. Sigmoid neurons are continuous, so AI can change weights to each neuron by calculating the direction of $- \nabla C$. AI learns that way, so continuity of neurons are important.

My question:

  1. Can human neurons be interpreted as a combination of perceptron neurons and sigmoid neurons? Human neurons have threshold values, which resembles perceptron neurons, and control the amount of nt, which resembles sigmoid neurons.
  2. If neurons can be interpreted as a combination of those two neurons, what makes the human brain and AI so different? Why do experts predict that AI will outsmart humans when (currently) the human brain can compute faster than AI?


Browse other questions tagged or ask your own question.