通过缪子反常磁矩在强度前沿寻找新物理

粒子物理学的标准模型在解释各种物理现象和几乎所有实验结果方面都取得了惊人的成功。目前最显著的差异之一是缪子的反常磁矩。2004年，美国BNL缪子g-2合作组所测得值比标准模型预测大于3倍标准方差。虽然许多学者提出了各种超越标准模型理论（如超对称粒子，暗光子，拟轴子粒子等等）来解释这个差异，但是解决这个偏差的关键仍然是更精确的测量值和理论预测值。费米实验室的缪子g-2合作组旨在以4倍以上的精确度，通过比较在同一个储存环磁场中的缪子和质子的进动频率之比来测量反常磁矩。任何在实验测量时间内的束流动态行为，或者储存环和探测器的不稳定行为，将会导致缪子进动频率的系统偏移。由于实验的系统误差的最终目标是千万分之一，降低这种系统偏移至关重要。本项目的目标是利用超级计算机设施和机器学习技术来构建实验的高精度仿真模型。 由于该实验现已收集到的事例是BNL的7倍，系统误差的估计已进入未知领域。因此，这种高保真且高速仿真对于分解各种系统效果至关重要。

The Standard Model (SM) of particle physics has been strikingly successful in explaining a wide range of physical phenomena and almost all experimental results. One of the most significant discrepancies at present is the anomalous magnetic moment of the muon (muon g-2), which was measured by the E821 collaboration at BNL in 2004 to be more than three-standard-deviation greater than the SM prediction. This persistent discrepancy between experiment and theory continues to fuel speculative models such as supersymmetry, dark photons, axion-like particles, and beyond. Hence, confirmation of the discrepancy with improved experimental and theoretical uncertainty is one of the priorities in the particle physics community. The Muon g-2 collaboration (E989) at Fermilab aims to measure muon g-2 with a precision goal of 140 parts per billion, a fourfold improvement over BNL's result. If the central values of the experimental result and the theoretical prediction remain unchanged, the discrepancy will exceed 7 standard deviations - a clear indication of beyond SM physics...E989 experiment measures AMM by comparing the ratio of the muon and proton precession frequencies in the same magnetic field of a storage ring. Any dynamic behavior of the muon beam or instability in the storage ring and detector components during the measurement cycles can shift the average muon polarization or beam distribution. Such a shift can introduce a time-dependent average phase shift, and if unaccounted for, will manifest itself as a systematic shift in the muon precession frequency. As the budget for the systematic uncertainty of E989 is 100 ppb, precise control of these effects are extremely crucial. In this project, I plan to utilize both supercomputer facilities and machine learning techniques to build a high-precision simulation model of the experiment. As the experiment has collected 7 times more muon decays than BNL, estimation of systematic effects has now gone into uncharted territory and therefore such high-fidelity and high-speed simulation is extremely crucial into disentangling various systematic effects.