Report title: An exact penalty approach for optimization with nonnegative orthogonality constraints
Reporter: Associate Professor Jiang Bo Nanjing Normal University
Reporting time: 9:40-10:20 am on January 4, 2021
Report location:Tencent Conference ID: 320 247 940
Conference password: 9999
School contact: Li Xinxin xinxinli@jlu.edu.cn
Report summary:
Optimization with nonnegative orthogonality constraints has wide applications in machine learning and data sciences. It is NP-hard due to some combinatorial properties of the constraints. In this talk, we shall discuss an exact penalty approach for solving the considered problems. The penalty model can recover the solution if the penalty parameter is sufficiently large other than going to infinity. We establish the convergence of the penalty method under some weak and standard assumptions Extensive numerical results on the orthogonal nonnegative matrix factorization problem and the K-indicators model show the effectiveness of our proposed approaches.
Speaker's profile:
Jiang Bo, associate professor of the School of Mathematical Sciences, Nanjing Normal University, a master tutor, and a young director of the Mathematical Planning Branch of the Chinese Operational Research Society. Selected in the 3rd China Association for Science and Technology "Young Talents Support Project". The main research direction is nonlinear optimization algorithms and theories, especially optimization problems with orthogonal constraints and their applications. They have been optimized and information in Mathematical Programming, SIAM Journal on Optimization, SIAM Journal on Scientific Computing, IEEE Transactions on Image Processing, etc. 7 papers published in top journals. Currently, he presides over 1 general project of the National Natural Science Foundation of China, 1 project of the National Natural Science Foundation of China Youth Project, and 1 project of the Jiangsu Provincial Youth Fund Project.
