Neural Networks – Garbage in, garbage out

I came across this article and I am somewhat concerned about how much we are potentially relying on neural nets and their “supposed” efficacy.


One comment on “Neural Networks – Garbage in, garbage out”

  1. Very good article! I love that it discusses the foundation for adequate machine learning. “Undetected biases in the input datasets may produce unintended results.”

    I have encountered this same problem in an industrial setting. Most people relate data-sets to absolutes (i.e. black & white), but many of these data-sets have influences, such as peoples individual perception, or emotion at the time of capture.

    Stanford teachings lessons on Decision Analysis which assigns probabilities on peoples personal beliefs; I would love to see some machine learning algorithms based on these probabilities that people assign and see what the numbers show.


    Users who have LIKED this comment:

    • avatar

Comments are closed.