报告题目: An Optimization View Point of Deep Learning: From Theory to Applications (三)
报告人： 贾西西 副研究员 （西安电子科技大学）
报告地点： 腾讯会议 330 850 001
Deep learning has become a milestone in a lot of applications ranging from computer vision, to natural langue processing, to auto drive. In this talk, I will give a general analysis of deep learning from the optimization point of view. To be more specific, I will show that how does each of the three ingredients in deep learning — training data, network architecture and optimization algorithms --affect the overall learning and generalization. Meanwhile, I will expose some tricks that really help improving the training efficiency.
Xixi Jia was graduated from Xidian University, and currently he serves at the school of mathematics and statistics of Xidian University. He also works as a research assistant at the Hong Kong Polytechnic University. His research interests include deep learning, differential equation, optimization and low-level vision.