Language Technologies Institute
School of Computer Science
Carnegie Mellon University
5000 Forbes Ave
Pittsburgh, PA 15213
Email: whlin [at] cs [dot] cmu [dot] edu
Office: SCFG 417 RM 301
Office Phone: +1-412-268-3119
Wei-Hao Lin (pronounced WAY-how-LIN) was a Ph.D. student in the Language Technologies Institute, School of Computer Science, Carnegie Mellon University. He defended his PhD thesis in October 2008. He works at Google now.
His Ph.D. thesis focuses on developing computer programs that can automatically identify ideological perspectives expressed in text and video, for example, identifying the Democratic or Republican perspective in the articles about the United States politics. You can find more details about automatic ideological analysis in his thesis proposal, web demos, and the data he has collected. He was chosen as one of the 2007 IBM emerging leaders in Multimedia for his work on identifying ideological perspectives in television news. He was a finalist in the 2007 Knight News Challenge for a proposal of automatically identifying biased news articles.
With his advisor Alex Hauptmann and colleagues in the Informedia project, he has worked on many problems in the field of multimedia, particularly developing classifiers that statistically combine evidence from multiple modalities (e.g., text, image, and audio). His "sports" classifier performed best in the 2005 TREC Video Track, and his "boat/ship" classifier performed best in 2004. He won the championship in the Physiological Data Modeling Contest in 2004 for his human activity classifier based on signals from wearable physiological sensors. He was co-author of a paper nominated for Best Paper Award at CIVR in 2007. He developed a video summarization system that achieved the best inclusion scores in the 2008 TRECVID BBC rushes summarization task.