Turns out that the QCBM can be interpreted as a Fourier model. We give the dequantization conditions and unify the generative surrogate methods into one same framework. We also pinpoint the sources of discrepancy when training on classical and deploying on quantum. scirate.com/arxiv/2511.0...