A Reversible Continuous-Variable Photonic Memory Architecture Based on Displacement-Evolved Coherent-State Dynamics

This paper proposes a novel continuous-variable (CV) photonic memory architecture designed to store and track the temporal evolution of information. Unlike traditional quantum memories that focus on preserving a single static state, this framework treats memory as a dynamic trajectory in phase space. The system uses a time-multiplexed optical loop to store a fixed number of latent qumodes, which are updated via electro-optic displacement operations. A classical digital-twin controller monitors these states through sparse homodyne measurements, recording displacement histories and metadata to enable the reconstruction of historical memory states.

The architecture demonstrates that complex information—such as temporal data streams or machine-learning latent representations—can be encoded into a small, fixed set of coherent-state pulses. By recording the sequence of displacement operations, the system allows for "rollback retrieval," where previous states can be reconstructed by applying inverse displacements. Numerical simulations indicate that retrieval fidelity remains high in low-noise regimes and degrades predictably as noise accumulates. The use of image entropy as a retrieval descriptor further enhances the efficiency of indexing and navigating historical data within the memory.

Photonic platforms are highly desirable for high-bandwidth, low-noise information processing but have historically lacked the ability to buffer and track evolving signals. This architecture provides a scalable, metadata-aware approach that decouples the number of physical qumodes from the resolution of the input data. By combining the speed of photonic hardware with the indexing capabilities of a classical digital twin, this framework offers a promising foundation for future systems that require the storage and reconstruction of dynamically changing information, such as scientific simulations and real-time data streams.