Report title: Adaptive Regularized Method for Optimization on Riemannian Manifold
Reporter: Researcher WEN Zaiwen (Beijing International Mathematics Research Center, Peking University)
Report time: October 8, 2016 AM 8:30-9:30
Location: A1101#room, Innovation Park Building
Contact: Prof. ZHANG Liwei (tel: 84708351-8118)
Abstract: Optimization on Riemannian manifold widely arises in eigenvalue computation, density functional theory, Bose-Einstein condensates, image and signal processing. We propose a second-order type approximation to the original problem and apply a first-order type method solve it. Global convergence to the first-order optimality conditions is established. Preliminary numerical experiments show that our method is promising.
Brief introduction to the reporter: Dr. WEN Zaiwen is a researcher of 100 Talents Program of Beijing International Center for Mathematical Research, Peking University. He received his PhD degree at Columbia University (Sept. 2009), his Master degree at Chinese Academy of Sciences (July. 2004) and Bachelor degree at Shanghai Jiao Tong University (July. 2001). He acquired the first Prize of the Science & Technology Award for Young and Middle-aged Talent, Operations Research Society of China (2012) and was the winner of "Education Ministry's New Century Excellent Talents Supporting Plan". He was supported by National Science Fund for Distinguished Young Scholars, The national natural science foundation (2013), and elected into the second batch of youth talents of "thousands of talents program" (2015). He was honored with The 14th Science & Technology award for Young and Middle-aged Talent (2016). His research focuses on Large-scale computing optimization and its application. He has already delivered more than 30 papers on International top magazines such as SIAM J. Optim., SIAM J. Sci. Comput., Math. Program. etc.