摘要
由一名新闻记者-机器人与机器学习每日新闻编辑-调查人员讨论人工智能的新发现。根据NewsRx记者从新墨西哥州洛斯阿拉莫斯发回的新闻报道，研究表明，“尽管二氧化铀E(UO2)是一种广泛使用的核燃料，燃料性能模型广泛依赖于材料行为的经验相关性，”利用UO2的历史OPE评级经验。考虑控制燃料性能过程的原子基础（如裂变气体释放和蠕变）的力学模型将能够更好地描述非原型AL条件下的燃料行为，例如在新的反应堆概念或改进的UO2燃料组合中。我们的新闻记者从洛斯阿拉莫斯国家实验室的研究中获得了一句话：“为此，分子动力学模拟是快速预测候选燃料物理性质的有力工具。然而，这些模拟的可靠性在很大程度上取决于原子力的准确性。传统上，这些力是用经典力场(FF)或密度泛函理论(DFT)计算的。虽然密度泛函理论相对精确，但计算成本很高，特别是对于F电子元素，如锕系元素。相比之下，经典的FFs计算效率较高，但精度较低。由于这些原因，我们报道了一种新的精确机器学习UO2原子间势(MLIP)，它以与经典FFs相似的低成本提供了高保真的DFT力再现。我们采用了一种主动学习方法，自主地增加DFT训练数据集来迭代地细化MLIP。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators discuss new findings in artificial intelligence. According to news reporting originating from Los Alamos , New Mexico, by NewsRx correspondents, research stated, "Despite uranium dioxid e (UO2) being a widely used nuclear fuel, fuel performance models rely extensive ly on empirical correlations of material behavior, leveraging the historical ope rating experience of UO2. Mechanistic models that consider an atomistic understa nding of the processes governing fuel performance (such as fission gas release a nd creep) will enable a better description of fuel behavior under non-prototypic al conditions such as in new reactor concepts or for modified UO2 fuel compositi ons." Our news journalists obtained a quote from the research from Los Alamos National Laboratory: "To this end, molecular dynamics simulation is a powerful tool for rapidly predicting physical properties of proposed fuel candidates. However, the reliability of these simulations depends largely on the accuracy of the atomic forces. Traditionally, these forces are computed using either a classical force field (FF) or density functional theory (DFT). While DFT is relatively accurate, the computational cost is burdensome, especially for f-electron elements, such as actinides. By contrast, classical FFs are computationally efficient but are l ess accurate. For these reasons, we report a new accurate machine learning inter atomic potential (MLIP) for UO2 that provides high-fidelity reproduction of DFT forces at a similar low cost to classical FFs. We employ an active learning appr oach that autonomously augments the DFT training data set to iteratively refine the MLIP."
NETL
NSTL