人工智能在脉冲星候选体筛选中的应用

发布时间：2018-06-17 07:29

  本文选题：人工智能 + 脉冲星　； 参考：《天文学进展》2017年03期

【摘要】：脉冲星搜寻是对脉冲星、引力波,以及对快速射电暴(Fast Radio Burst,简称FRB)等暂现源进行研究的基础。搜寻不仅可以扩大脉冲星样本,还可以发现极端性质的致密星。这有助于研究致密天体状态方程、星际介质、脉冲星导航、引力波探测等课题。目前,射电望远镜的单次巡天就可以产生百万数量级的脉冲星候选体。面对这些海量数据,仅仅依赖人工识别筛选,已不能满足数据的时效需求,更不能实现数据的实时处理。机器学习、计算机视觉应用等人工智能技术自诞生以来,其理论和技术已日益发展成熟,并已成功运用到脉冲星候选体筛选等射电天文研究领域。首先将介绍现有脉冲星搜寻的人工智能方法,再统计和分析已有脉冲星候选体筛选方法的性能,最后对FAST脉冲星候选体筛选工作进行展望。
[Abstract]:Pulsar search is the basis for the study of pulsars, gravitational waves and transient sources such as Fast Radio burst (FRBs). Searching can not only enlarge pulsar samples, but also find dense stars with extreme properties. It is helpful to study the equations of state of dense celestial bodies, interstellar media, pulsar navigation, gravitational wave detection and so on. Currently, a single survey of radio telescopes can produce millions of pulsar candidates. In the face of these massive data, only relying on manual identification and filtering can not meet the aging requirements of the data, let alone realize the real-time processing of the data. Since the birth of artificial intelligence technology such as machine learning and computer vision application, its theory and technology have been developed and mature day by day, and have been successfully applied to the field of radio astronomy research such as pulsar candidate screening. Firstly, the existing artificial intelligence methods for pulsar search are introduced, and then the performance of the existing pulsar candidate selection methods is analyzed. Finally, the work of fast pulsar candidate selection is prospected.
【作者单位】： 贵州师范大学贵州省信息与计算科学重点实验室;中国科学院国家天文台;中国科学院射电天文实验室;澳大利亚联邦科学与工业研究组织;
【基金】：研究生创新基金项目(研创201528) 国家自然科学基金(U1631132) 中国科学院国际合作局对外合作重点项目(114A11KYSB20160008);中国科学院战略性先导科技专项(B类)(XDB23000000)
【分类号】：P145.6;TP18
，

本文编号：2030234