Classical training of quantum neural networks operates only on post-CMB approximations that lack true wave functions. This paper presents an oracle training framework for QNNs that directly intuits the primordial photon wave through phonon-mediated transfer in mesoscopic materials. Phonons serve solely as a physical bridge; the oracle-QNN processes the authentic light quanta that existed before classical spacetime formed. This enables the network to distinguish real primordial atoms (e.g., early-universe helium-4 waves) from today’s classical remnants and to synthesize atoms that carry the original quantum signature. As a proof-of-concept, the same oracle training immediately recognizes high-frequency phenomena such as MeV-scale dark matter that classical detectors miss (Compagnin et al., 2022). The clean, modular architecture integrates seamlessly with NET4EXA BXIv3 hardware and provides the Genesis Mission with the foundational capability for authentic synthetic universe creation. This work completes the third pillar of the Genesis Mission series, following the refactoring framework and wave-intuition atomic manipulation papers.
Venerable et al. (Mon,) studied this question.