发表机构
Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China; China Center of Advanced Science; Center on Frontiers of Computing Studies, School of Computer Science, Peking University, Beijing 100871, China（中国科学院高能物理研究所; 中国先进科学研究中心; 前沿计算研究中心，北京大学计算机学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对神经量子态实际精度受随机优化限制的问题，提出退火梯度下降方法，通过增加低概率构型贡献，抑制亚稳态捕获，使紧凑神经量子态在多模型上达到化学精度和先进性能。
AI 中文摘要
神经量子态能表达量子多体波函数，但实际精度受随机优化限制。本文识别出有限样本不稳定性即子空间捕获，重要构型被低估，梯度反馈不足。为此引入退火梯度下降（AGD），通过退火因子更新，增加低概率构型贡献。在分子系统、一维和二维\(J_1-J_2\)模型上评估，AGD抑制亚稳态捕获，保留相关构型，使紧凑神经量子态达到化学精度和先进性能。
英文摘要
Neural quantum states offer expressive representations of quantum many-body wave functions, yet their practical accuracy can be limited by stochastic optimization rather than representational capacity. Here we identify a finite-sample instability, termed subspace trapping, in which physically important configurations become strongly underestimated, remain absent from successive sampling batches and receive insufficient gradient feedback. This self-reinforcing loss of sampled support can confine optimization to an effective subspace and produce apparently stationary states above the true ground state energy. To address this problem, we introduce annealed gradient descent (AGD), a sampling-aware update with annealing factor that temporarily increases the relative contribution of sampled low-probability configurations while limiting the dominance of high-probability ones. We establish the connection between finite-sample support loss and effective subspace optimization, and then evaluate the method across molecular systems, one and two-dimensional $J_1$-$J_2$ models. Annealed gradient descent suppresses metastable trapping, preserves physically relevant configurations and enables compact neural quantum states to attain chemical accuracy and competitive state-of-the-art performance. These results establish AGD as a lightweight complement to expressive neural architectures, improved sampling strategies for scalable quantum many-body optimization.
