SymQNet: Amortized Acquisition for Low-Latency Adaptive Hamiltonian Learning

SymQNet is a novel amortized reinforcement-learning framework designed for low-latency adaptive Hamiltonian learning in quantum devices. By learning a posterior-conditioned acquisition policy offline and executing a rapid policy forward pass online, it achieves significant reductions in decision latency—up to 72.6 times faster than traditional methods like bounded Fisher-information search and bounded two-step Bayesian active learning by disagreement on a twelve-qubit system. This advancement is critical for practitioners seeking efficient adaptive control in quantum applications, enabling practical implementation in scenarios requiring rapid decision-making.