In Quanvolutional Function works without using filters
could you explain how Quanvolutional function works and what is expected output?
Thanks for your question.
If the tutorial you’ve taken the screenshot from is not quite enough for you, the other thing we can recommend is the original paper from Max Henderson and coauthors, which obviously will go into much more details.
I could also suggest brushing up—if you’re not already familiar—on the original concept of a convolutional neural network from classical machine learning.
@nathan many thanks for your reply
I read the original paper, in the original paper, the authors use filters from 1 to 50 in Quanvolutional layer, and these filters not clear in quanv funtion.
My question , is quanv function in your tutorial (https://pennylane.ai/qml/demos/tutorial_quanvolution.html)
use these filters or not?
thanks in advance.
quanv mapps a 2 x 2 square made of 4 pixels into
a single pixel characterized by 4 channels (i.e. 4 parameters) corresponding to the 4 expectation values measured in the quantum circuit. So, if I am not wrong, I think that in the language of CNNs the function corresponds to a convolution with 4 filters, each having a kernel size of 2 x 2 and a stride of 2.
I apologize for the delay in reply.
Thank you very much for your explanation.
I could use it Rx gate for encoding instead of Ry
and many thanks.