摘要
基于贝叶斯神经网络预测了光中子反应的光中子反应道——(γ,n)和(γ,2n)的反应截面。优化器采用Torch.Adam,学习率设置为0.0001。在训练过程中,结合物理信息引导有效增强了贝叶斯神经网络对数据结构的捕获能力。优化后的贝叶斯神经网络算法显著缩短了训练时间,单个模型的训练时间不超过7 min,为大量模型测试提供了保障。模型训练完成后,分析了B2、B3和B4模型(隐含层数量分别为2~4)的训练集预测结果,发现随着神经网络隐含层数的增加,模型对数据结构的捕捉能力逐渐增强,且在相同训练次数下表现更为突出。对于训练集中的^(89)Y和^(159)Tb,所有模型均能较好地描述单、双巨共振峰核的(γ,n)和(γ,2n)反应截面,并准确给出巨共振峰的峰位、峰宽和峰高。随后,使用不同模型对197Au的(γ,n)和(γ,2n)反应截面进行了预测,并与实验数据进行了对比。结果表明:与训练集相似,隐含层数量最多的B4模型预测结果最佳。最后,对175Lu的预测分析显示,B4模型可以精确地给出双峰核的峰位和峰相对高低。随着隐含层的增加,模型对训练集的拟合效果和泛化能力均得到了显著提升。未来,基于物理信息的贝叶斯神经网络有望提供更多可靠的光核反应截面数据,为解决实验室间数据分歧提供有力支持。
[Background]The origin of elements is a significant research topic in nuclear physics and astrophysics.Some heavy nuclei are produced through photonuclear reactions,known as p-nuclei.The study of photonuclear reactions plays a crucial role in understanding the origins of elements.The existing data on photoneutron reactions have significant discrepancies.It is well-known that the(γ,n)reaction cross-sections from Saclay are higher than those from Livermore,while conversely,the(γ,2n)cross-sections from Livermore are higher than those from Saclay.To resolve these divergences,we need to either remeasure these data or evaluate them based on theoretical models.[Purpose]This study aims to predict the photoneutron reaction cross-sections,specifically(γ,n)and(γ,2n)reactions,using a Bayesian neural network(BNN).The goal is to develop a physics-informed BNN model that improves the accuracy of photoneutron reaction predictions and resolves divergence in experimental data from different laboratories.[Methods]A physics-informed Bayesian Neural Network(PIBNN)model was constructed using PyTorch,designed to predict the photoneutron reaction cross-sections.The network was trained with a consistent dataset from Livermore's photoneutron experimental data,incorporating physics-informed such as crosssections are zero,below reaction thresholds.The B2,B3 and B4 network architectures include various hidden layers(2,3,and 4 layers),with an Adam optimizer and a learning rate of 0.0001.[Results]As the number of hidden layers increases,the model's description of the training set improves with the same number of training iterations.Among them,the B4 model not only effectively reproduces the single and double giant dipole resonance(GDR)peak structures of the(γ,n)reaction channel in the training set,but also accurately captures the magnitude of the(γ,2n)cross-section.The physics-informed incorporated into the training set,particularly the inclusion of zero cross-sections below reaction thresholds,improved the model's accuracy in predicting the cross-section near the threshold and ensuring that cross-sections approach zero at high energies.The predictions of the(γ,n)and(γ,2n)reaction cross sections for ^(175)Lu by the three models are compared with the Saclay experimental data.The B4 model accurately provides the position and relative heights of the double-peak structure,reflecting the inherent systematics of the training set.By predicting the(γ,n)and(γ,2n)reaction cross-sections for ^(197)Au and ^(175)Lu,it has been validated that the trained physics-informed Bayesian neural network model possesses generalization ability.[Conclusions]Based on the physics-informed Bayesian neural network,the model can effectively learn the(γ,n)and(γ,2n)reaction cross-sections,reproducing the data in the training set and predicting cross-section data outside the training set.Furthermore,as the number of hidden layers increases,the model's learning ability gradually improves.In the future,the trained B4 model can be used to predict reliable photoneutron reaction cross-sections,resolve data discrepancies between different laboratories.
作者
孙乾坤
张岳
郝子锐
王宏伟
范功涛
许杭华
刘龙祥
陈开杰
金晟
王振伟
徐孟轲
王向飞
SUN Qiankun;ZHANG Yue;HAO Zirui;WANG Hongwei;FAN Gongtao;XU Hanghua;LIU Longxiang;CHEN Kaijie;JIN Sheng;WANG Zhenwei;XU Mengke;WANG Xiangfei(Shanghai Institute of Applied Physics,Chinese Academy of Sciences,Shanghai 201800,China;University of Chinese Academy of Sciences,Beijing 100049,China;Shanghai Advanced Research Institute,Chinese Academy of Sciences,Shanghai 201210,China;ShanghaiTech University,Shanghai 201210,China)
出处
《核技术》
北大核心
2025年第5期1-10,共10页
Nuclear Techniques
基金
国家重点研发计划(No.2022YFA1602404,No.2023YFA1606901)
国家自然科学基金(No.12388102,No.U2441221,No.12275338)
核数据重点实验室基础项目(No.JCKY2022201C152)资助。
关键词
光中子反应
贝叶斯神经网络
机器学习
γ源
上海激光电子γ源
Photoneutron reaction
Bayesian neural network
Machine learning
Gamma source
SLEGS(Shanghai Laser Electron Gamma Source)
