two questions about conventions:
1. Do we always assume the prototype neurons have values either +1 or -1 ?
2. Is the formula for scaled by 1/N or not ? since in the slides page 13 it is scaled but in page 8 it is not.
Hi !
Thank you for your question.
1) this week you had the first course on Hopfield Networks, indeed so far you have seen only -1,1 neurons. Though there exists other possiblities. You will have an overview next week on the generalized hopfield models. Comeback to me if you still have question after next week.
2) The scaling factor is not so important, its a convention. Most of the time it is scaled. You can convince yourself that It does not change the dynamics of the Hopfield networks we have seen so far.
I hope I answered your questions,
Martin.
Follow the pulses of EPFL on social networks
© 2023 EPFL, all rights reserved