摘要
由一名新闻记者兼机器人与机器学习的新闻编辑每日新闻-调查人员发布了关于马学习的新报告。根据NewsRx编辑在宾夕法尼亚大学公园的新闻报道，研究表明，“γ射线光谱学在核科学、核安全和环境监测中是OL必不可少的。然而，由于存在低计数、多来源和动态背景，解释光谱数据面临挑战。”这项研究的财政支持者包括国防威胁减少局(DTRA)，作为电离辐射相互作用的一部分，物质大学默多克联盟(IIRM-URA)。我们的新闻记者引用了宾夕法尼亚州立大学（宾州州立大学）的研究，“为了解决这些问题，开发了一种利用机器学习技术进行γ射线谱分析的新的特征驱动分析方法。该方法利用一系列随机森林模型进行分布(ID)多标记分类，该方法在不同光谱参数下,包括采集时间、源数、源能量、背景成分等,定量评价了该方法的性能,将采集时间从1s增加到100s,提高了多标签分类的性能。对于CLLBC手持探测器，采集50 s后，22个源的F1分数>=0.9.特性驱动的分析应用程序Roach在处理复杂源混合物时也表现出稳健性。此外，它还为OOD检测提供了上下文能量信息。这里给出的结果强调了该方法的可解释性。在光谱特征和底层物理之间建立了清晰的联系。此外，该方法有效地区分了不同IDγ射线源重叠的光谱特征，提高了基于机器学习的γ射线光谱分析中的人的可靠性。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News – Investigators publish new report on Ma chine Learning. According to news reporting out of University Park, Pennsylvania , by NewsRx editors, research stated, “Gamma-ray spectroscopy is an essential to ol in nuclear science, nuclear security, and environmental monitoring. However, challenges arise in interpreting spectral data due to the presence of low counts , multiple sources, and dynamic backgrounds.” Financial supporters for this research include Defense Threat Reduction Agency ( DTRA) as part of the Interaction of Ionizing Radiation, Matter University Resear ch Alliance (IIRM-URA). Our news journalists obtained a quote from the research from Pennsylvania State University (Penn State), “To address these issues, a novel feature-driven analyt ical approach for gamma-ray spectral analysis using machine-learning techniques is developed. The method utilizes a series of random forest models for in-distri bution (ID) multi-label classification, and the model-derived feature importance values to guide the out-of-distribution (OOD) detection task. The performance o f this approach is quantitatively evaluated across various spectral parameters, including acquisition time, number of sources, energy of an OOD source, and back ground composition. Increasing the acquisition time from 1 s to 100 s leads to i mproved performance for multi-label classification, with 22 sources achieving F1 -scores >= 0.9 after 50 s acquisitions for a CLLBC handh eld detector and a standoff distance of 30 cm. The feature-driven analytical app roach also demonstrates robustness when handling complex source mixtures. Furthe rmore, it provides contextual energetic information for OOD detection. The resul ts presented here highlight the interpretability of the approach, establishing c lear links between the spectral features and underlying physics. Moreover, the a pproach effectively distinguishes overlapping spectral signatures of different I D gamma-ray sources, enhancing human reliability in machine learning-based gamma -ray spectral analysis.”
NETL
NSTL