主讲人：Dr. Jun Wang (IBM T.J. Watson Research Center)
时间：2014 年 6 月 27 日（周五）下午 2:30-3:30
地点：软件楼 105 IBM 会议室
联系人：姜育刚（ygj@fudan.edu.cn）
Abstract:
Fueled by the ever-increasing amount of data, there is an emerging need to develop efficient models and algorithms to handle the scalability, heterogeneity, and privacy of big data. However, many real world applications have a nontrivial gap between theory and practice due to several factors, e.g., the semantic gap and insufficient labeling. In this talk, I will discuss several topics related to large scale applications. First, we have leveraged learning algorithms to design efficient hashing functions for efficient indexing, ranging from unsupervised to semi-supervised to supervised hashing techniques. Second, I will talk about several prediction algorithms that explore the heterogeneity of data, including heterogeneous random walk and heterogeneous indexing algorithm. Finally, I will briefly discuss the privacy preserved learning techniques for data sharing and model training.
Bio:
Jun Wang is a Research Staff Member in the data science group at IBM T.J. Watson Research Center in Yorktown Heights, New York (2010-Present). He received the PhD degree from Columbia University, the MS degree from Tsinghua University, and the BS degree from Shanghai Jiaotong University. He is also an adjunct assistant professor at Columbia University, teaching the graduate course “Visual Recognition and Search.” He worked as a research intern at Google Research in 2009, and as a research assistant at Harvard Medical School, Harvard University, in 2006. He has been the recipient of several awards and scholarships, including the Outstanding Technical Achievement Award from IBM in 2013, the Jury award for his PhD thesis in 2011, the IBM T.J. Watson Emerging Leader in Multimedia Award 2009, the Google Global Intern Scholarship in 2009, and a Chinese government scholarship for outstanding self-financed students abroad in 2009. His research interests include machine learning, data mining, computer vision and healthcare analytics.
