报告人：彭思愿

题目：非负矩阵分解理论研究

时间：2022年8月02日（周二）19:30

会议链接：腾讯线上会议码：726-878-341

内容摘要：

Nonnegative matrix factorization (NMF) is one of the most powerful dimensionality reduction techniques for the ****ysis of high-dimensional data, and it has been successfully used in the fields of machine learning and data mining due to the simplicity and intuitive decomposition. The past decade has witnessed a rapid development of the NMF techniques, which develops many new NMF methods to solve different practical issues. However, traditional NMF methods still suffer from the data with non-Gaussian noise and outliers, and fail to make full use of limited supervised information for improving the performance of the NMF methods. In recent years, the information theoretical learning and semi-supervised learning techniques have been incorporated into NMF model to overcome the above mentioned problems.

Some novel robust and semi-supervised NMF methods will be introduced.

个人介绍：

   彭思愿，广东工业大学信息工程学院讲师。主要研究方向为非负矩阵分解理论、麻醉深度预测以及信息理论学习等，已在Neural Network, Pattern Recognition, Information Sciences, Knowledge Based System等国际期刊上发表SCI检索论文12余篇；担任国际期刊Processes客座编辑，主持教育部重点实验室开放课题1项。