Variational Quantum Transduction

Quantum transducers play a central role in the quantum internet by enabling efficient transfer of signals across different frequency domains. Beyond material and device engineering, protocol design is another key lever for improving transduction performance. Recent advances in protocol design include schemes based on teleportation mechanisms, Gottesman–Kitaev–Preskill (GKP) encoding, entanglement assistance, and related techniques. To move beyond heuristic approaches, we propose a variational quantum transduction (VQT) paradigm that leverages tools from near-term quantum computing to aid transducer design. By directly optimizing the quantum information rate, we identify protocols that surpass all existing methods within their respective classes. For non-adaptive protocols, VQT exceeds the envelope of GKP-based approaches [Phys. Rev. X 15, 021003 (2025)] and entanglement-assisted approaches [Optica Quantum 2, 475 (2024)], demonstrating a substantial advantage. For adaptive protocols, VQT outperforms the previous Gaussian adaptive proposal [Phys. Rev. Lett. 120, 020502 (2018)], while unexpectedly revealing the near-optimality of Gaussian adaptive schemes. We also clarify the resources favored by VQT: entangled GKP states in the non-adaptive class and squeezed-thermal states in the adaptive class. With the advent of universal quantum control across multiple platforms, our results pave the way toward achieving optimal quantum transduction.