中文

杨明 副教授

简介

当前位置: 中文主页 > 科学研究 > 专利 > 简介
本人的研究方向聚焦于新一代人工智能技术的前沿领域,主要研究基于低秩模型的高效处理高维数据的方法,致力于解决大数据产业中常见的数据结构复杂、带强噪声、有损毁或缺失等问题,以此克服传统机器学习方法失效的局限性。在多年的研究过程中,本人采用基于低秩三维数组的非凸优化方法,成功攻克了多视数据子空间聚类、高维数据的补全、三维图像显著区域检测以及视频前景背景分离等基础性难题。本人的研究成果在国内外产生了一定的影响,并得到了同行专家的高度评价。相关研究成果已发表于 IEEE T-SP (一作)、SIAM Journal on Imaging Sciences (一作)Neural Networks(通讯)、以及Pattern Recognition (一作+通讯)等权威国际期刊,涉及模式识别和机器学习领域。此外,本人还与多位合作者合作在多个JCR一区刊物上发表了10多篇论文,其中包括Information Sciences、IEEE汇刊T-CYB、T-MM 、T-KDE、T-IP等,被Google Scholar引用近200次。本人的研究成果得到了来自中、美、比利时等多个国家的引用和正面评价。 我是《Frontiers》编辑委员会的审稿编辑,专门负责《计算与数据科学数学》的编辑工作(这是《应用数学与统计前沿》的专栏部分)。

My research focuses on the cutting-edge domains of the next-generation artificial intelligence technologies. I mainly study efficient methods for processing high-dimensional data based on low-rank models. My work is dedicated to addressing prevalent issues in the big data industry, such as complex data structures, strong noise, damaged or missing data, thereby overcoming the limitations of traditional machine learning methods. Throughout years of research, I have employed non-convex optimization methods based on low-rank three-dimensional arrays, successfully tackling fundamental challenges such as multi-view data subspace clustering, high-dimensional data completion, three-dimensional image salient region detection, and video foreground-background separation. My research achievements have made a significant impact both domestically and internationally, receiving high praise from peers in the field. Pertinent research results have been published in authoritative international journals such as IEEE T-SP (1 as the first author), SIAM Journal on Imaging Sciences (1 as the first author), and Pattern Recognition (1 as the first author and 1 as the corresponding author), covering areas like pattern recognition and machine learning. Additionally, I have collaborated with several co-authors to publish over 10 papers in JCR Q1 journals, including Neural Networks, Information Sciences, IEEE Transactions on Cybernetics (T-CYB), T-MM, T-KDE, T-IP, among others. These works have been cited nearly 200 times on Google Scholar. My research has been positively recognized and cited by scholars from countries including China, the United States, and Belgium.