I have similar question to the following one.

Could an AI be built to learn based of interaction with a human?

What category of NN is the situation when the network learns from users feedback? The NN digests data and pass through. Than gives answer and a teacher review it by giving a score: 0.0 - failed, 1.0 - valid (or similar). I cannot imagine, how the user's answer should be entered to the network to train it.

  • $\begingroup$ you are referring to RLHF $\endgroup$
    – Alberto
    Commented Apr 29 at 15:11


You must log in to answer this question.

Browse other questions tagged .