Hardware-Efficient Large-Scale Universal Linear Transformations for Optical Modes in the Synthetic Time Dimension

Abstract: Recent progress in photonic information processing has generated strong interest in scalable and dynamically reconfigurable photonic circuitry. Conventional approaches based on spatial interferometer meshes face a fundamental scaling bottleneck, requiring a number of components that grows quadratically with system size. Here, we introduce a hardware-efficient time-domain photonic processor that achieves at least an exponential reduction in component count for implementing arbitrary linear transformations. Our design leverages the favorable scaling properties of the synthetic time dimension by encoding information in time-binned modes and processing them in parallel using recursive switchable short and long-range coupling. The dynamic connectivity of our processor enables systematic pruning of circuit depth, which minimizes optical loss while maintaining all-to-all connectivity. We benchmark our platform on the task of boosted Bell state measurements - a critical component in linear optical quantum computing, and demonstrate that the architecture surpasses the thresholds required for universal cluster-state quantum computation under realistic hardware parameters. We link the performance of our device to the geometric nature of multi-photon transport and show that, contrary to the expectation that redundant faulty hardware degrades performance, localization effects may contribute to improved robustness against coherent errors. Our results establish a practical pathway toward near-term, scalable, and reconfigurable photonic processors for quantum computation and simulation in the synthetic time dimension.