One can train binary NN via backprop by using qbits. Samples are taken, and backprop can be done as if the result was X% 1 and (1-X)% 0. It can be noisy and slo

One can train binary NN via backprop by using qbits. Samples are taken, and backprop can be done as if the result was X% 1 and (1-X)% 0. It can be noisy and slow but it is a way to do it.