Simge is programming a neural network to identify proteins in the human body. She will give the computer millions of examples of mass spectrometer readings and ask it to predict which proteins are present in the reading. She will then use supervised learning with backpropagation to tell the computer which of its predictions are correct and which are incorrect. What problem will Simge most likely encounter if she uses supervised learning to train the computer for this?

Question $9$ options:

a$)$

The computer might change its own programming, make itself more intelligent than Simge, and cause the Singularity.

b$)$

The computer will not be flexible enough to generalize its learning to new combinations of proteins, even if they are only slightly different from the ones in the training set.

c$)$

The computer will create its own categories for proteins which may not correspond to the existing categories that a human would understand.

d$)$

It will be very labor intensive to create the training data because each of the millions of example mass spectrometer readings will have to be manually labeled by a human.