I've been reading non-local neural networks as explained in the original paper. My understanding is that they solve the restrained reception of local filters. I see how they are different from convolutions and fully connected networks.
How do they relate to attention (specifically self-attention)? How do they integrate this attention?