I continue to try to understand how a human brain works and how we can use it to set up AI which would copy the behavior of a actual person. A few days ago I uploaded a question where I described a process of function of the brain of such a AI. However I have changed my mind on some things.What if let's say we have a correlation function and a neural network. My new idea is that first the correlation function will look for patterns between the input and some stored data or memories or logic. Then if the correlation coefficient is greater than some number let's say 0.7 then the neural network will process those data and do what a neural network does. This will ensure that we would have a dynamic pattern recognition process and that the AI could still learn from new experiences, although it would be able to understand the relevance of a input with a understood pattern(memory or logic). Is there such a mode in Machine Learning?

  • $\begingroup$ Hi. This question is a bit better than last, but I think it is still hard to join the dots well enough to figure out what you are asking, and what a useful answer could look like. I don't want to discourage you though, if you have a sincere interest in the subject. I think we could try to work on refining the question so that it still explores at least a part of your idea, and works somehow on the site. I'll have a think about that and comment further when I have more time $\endgroup$ Sep 27, 2023 at 7:52
  • $\begingroup$ I think we probably need to open a chat room, but as a start, let's look at your first statement. "I continue to try to understand how a human brain works" - could you explain, briefly what you are doing to increase your understanding? What material are you watching/reading that is relevant to understanding how the human brain works, and which parts of that material have inspired this question? $\endgroup$ Sep 27, 2023 at 8:14


You must log in to answer this question.