On Friday, January 31st, the Department of Information and Library Science hosted its first colloquium of 2025, featuring Dr. Siqi Wu’s presentation, “Towards Transparent Social Computing Systems.” This talk presented two studies, which both focused on the ways that people can and do use social media. By studying social media, Dr. Wu hopes to work towards building a social system that is fair, safe, and responsible.
The first study was a large-scale measurement study, meant to identify the patterns of news consumption at different places all around the world. Often, reports on media bias are editor-curated and they don’t include news media that could provide useful insights, such as news media from non-English-speaking countries. This study is useful for several different reasons: it can inform media outlets of their audience base and effectiveness, assist research organizations in understanding media audience, and help individuals reflect on their personal media consumption.
The second study was meant to deduce the actual usefulness of YouTube’s different functions for removing unwanted video recommendations. The study used “sock puppet” agents and two phases to test YouTube’s different functionalities. The first phase was the “stain phase,” where the agent watched numerous videos on an assigned topic. The second phase was the “scrub phase,” where the agent then attempted to remove recommendations that were related to their assigned topic. The study found that the “Not interested” button was the most effective in reducing unwanted recommendations; it had an average removal rate of 88% across all the assigned topics that were tested. In addition, the study also surveyed adult YouTube users about their awareness of the buttons, how much they use them, and their perceived effectiveness. The survey found that 44% of the users were unaware of the “Not interested” button. The colloquium was well attended, and afterwards Dr. Wu received many questions from graduate students and faculty!
Dr. Siqi Wu is an assistant professor of Information and Library Science in the Luddy School of Informatics, Computing, and Engineering at Indiana University Bloomington. His research aims to (i) advance our understanding of contemporary social phenomena via large-scale measurements and (ii) design future social computing systems via policies and interventions driven by data. He is the recipient of a best paper honorable mention award at ACM CSCW, a spotlight paper award at AAAI ICWSM, and a Google PhD fellowship. He serves as a senior PC member for ICWSM. Previously, he was a postdoc research fellow in the Center for Social Media Responsibility at the University of Michigan School of Information. He earned his Ph.D. (Computer Science) from the Australian National University, M.S. (Information Technology) from the University of Melbourne, and B.E. (Electronics Engineering) from Tianjin University.