AI in Multimedia Advertising
Wei Liu, Distinguished Scientist
Dr. Wei Liu is currently a Distinguished Scientist of Tencent, China. He takes the director of Ads Multimedia AI at Tencent Data Platform, and has taken the director of Computer Vision at Tencent AI Lab. In the meantime, Dr. Liu is a Guest Full Professor in the School of Computer Science, Fudan University, China and holds an Adjunct Faculty position in the School of Science and Engineering, the Chinese University of Hong Kong-Shenzhen, China. Prior to joining Tencent, he received the Ph.D. degree in EECS from Columbia University, USA, and was a research scientist and staff member of IBM T. J. Watson Research Center, USA. Dr. Liu has long been devoted to fundamental research and technology development in core fields of AI. His research works win numerous awards and honors, e.g., the 2011 Facebook Fellowship, the 2012 IBM Josef Raviv Memorial Postdoctoral Fellowship, the 2013 Jury Award for Best Thesis of Columbia University, the 2016 and 2017 SIGIR Best Paper Award Honorable Mentions, the 2018 “AI’s 10 To Watch” Award, the 2020-2021 World’s Top 2% Scientists honor, etc. Dr. Liu is an Associate/Action Editor of internationally leading AI journals IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), IEEE Transactions on Neural Networks and Learning Systems (TNNLS), IEEE Intelligent Systems, and Transactions on Machine Learning Research (TMLR), and is an Area Chair of top-tier computer science and AI conferences NeurIPS, ICML, CVPR, ICCV, IJCAI, and AAAI. Dr. Liu is a Fellow of the IAPR (International Association for Pattern Recognition), AAIA (Asia-Pacific Artificial Intelligence Association), IMA (Institute of Mathematics and its Applications), BCS (British Computer Society), and RSA (Royal Society of Arts), and is an Elected Member of the ISI (International Statistical Institute).
Abstract: In this talk, I plan to present core AI problems that contemporary online advertising in China encounters. Multimedia contents are widespread in Chinese Internet advertisements, which makes us to research and develop powerful deep learning techniques to understand, generate, and distribute those contents in a more precise, high-fidelity, and rapid manner. Our developed deep learning techniques include image/video classification, image/video captioning, image/video hashing, video segmentation and summarization, cross-modal retrieval, image segmentation and inpainting, 3D estimation from image, neural rendering, etc.