Abstract: As an important transitional technology before the maturity of fault-tolerant quantum computing, quantum simulation has already demonstrated significant practical quantum advantages and holds broad application prospects in fields such as condensed matter physics and new materials research. Due to the immaturity of large-scale quantum error correction techniques, the main challenge in quantum simulation lies in the coherent manipulation and characterization of large-scale quantum states. Effectively addressing this issue will help improve the simulation accuracy and computational reliability of quantum simulation systems. To tackle the manipulation and characterization of large-scale qubit systems, we propose a verification scheme for quantum simulation evolution system based on deep learning techniques. This scheme distinguishes between disordered and ordered states of a quantum system to realize state identification and property estimation. This deep learning-based scheme effectively verifies quantum simulators without relying on quantum state tomography, thereby establishing a robust foundation for the verification of future large-scale quantum simulation platforms.

Key words: quantum computing, quantum simulation, deep learning, randomized quantum measurement, quantum simulator verification