Large-Scale Analysis for Connected Media
Add to Google Calendar
We are living in a revolutionary age, witnessing the next-generation digital and social media emerged in astounding volume and rich formats. The rapidly grown, efficiently delivered, densely connected and incrementally socialized media content has fundamentally reshaped the ways for human to learn about this world, express their thoughts and interact with one another. They have also brought to computer scientists unprecedented opportunities and challenges in terms of computational analysis. In this talk, I will cover some of my recent research topics on analyzing and understanding large-scale media data with rich connections, allowing them to be served for users in a more effective and efficient way. I will first describe a large-scale visual concept detection algorithm for image and video understanding called model-shared subspace boosting. It merges multi-task learning and LogitBoost to automatically exploit sharing patterns across semantic label space, such as "car" and "road" , and hence significantly improve the concept detection performance. I will then present a multi-media retrieval approach based on hierarchical Bayesian models to discover the common latent mixing structure on heterogeneous media sources, including image retrieval, audio / speech retrieval, motion detection and visual concepts, followed by adaptively combining these retrieval results for unseen queries. Finally, I will also introduce a MapReduce-based distributed learning algorithm and a new offline job scheduling scheme on the Hadoop cloud computing platform to enable terabyte-scale analysis on real-world media collections.
Dr. Rong Yan is a currently Research Scientist in Facebook. He was a Research Staff Member in the IBM T. J. Watson Research Center from 2006 to 2009. Dr. Yan received his M.Sc. (2004) and Ph.D. (2006) degree from Carnegie Mellon University's School of Computer Science. His research interests include digital / social media analysis, large-scale machine learning / data mining, information retrieval and computer vision. Dr. Yan received the Best Paper Runner-Up awards in ACM Multimedia 2004 and ACM CIVR 2007. He has received the IBM Research External Recognition Award in 2007. He led the technical efforts for building the face detection service in Facebook, and also designed the automatic video retrieval system that achieves the best performance in the world-wide TRECVID evaluation in 2003 / 2005. Dr. Yan has authored or co-authored 5 book chapters and more than 60 international conference and journal papers. Dr. Yan has served or is serving as co-chairs for 10 conferences / workshops and as a Program Committee member in more than 40 ACM / IEEE conferences. He is an expert reviewer for more than 10 international journals. He has served in the NSF proposal review panel and as reviewers for several other research councils. Dr. Yan gives tutorials and guest lectures at several major conferences and universities.