I saw recently a few posts and papers about hybrid quantum-classical CNN. First a quantum circuit is designed for convolutional and pooling layer and then a classical fully connected layer is used as dense layer. How is this possible? The output of quantum circuit is still a quantum state after applying convolutional layer and pooling layer (and before applying dense layer). How can a quantum state be fed into a classical layer (classical neural network)?
Hey @sassan_moradi,
How can a quantum state be fed into a classical layer (classical neural network)?
This is kinda magical, but so long as you abstract away the fact that we’re dealing with a quantum circuit and simply look at it like a black box that maps inputs to outputs, it’s entirely possible to connect a quantum layer to a classical one 
.
Hey @isaacdevlugt, that’s very interesting. Are there any examples available that I could explore to better understand it?
We have a great demo here! Quanvolutional Neural Networks | PennyLane Demos
Hey @isaacdevlugt. Many thanks again. 