Mixing quantum annealing and classical networks beats either alone

Instead of jointly optimizing quantum and classical components (a notoriously difficult problem), this approach freezes a quantum-annealed state and uses it as a boundary condition for a classical tensor network that is then optimized on its own. Tested on the transverse-field Ising model, the hybrid beats the raw quantum simulation in accuracy without adding circuit depth. The trick sidesteps the fragile variational quantum optimization step that has hampered most hybrid schemes.