Please activate JavaScript!
Please install Adobe Flash Player, click here for download

uni'wissen 01-2016_ENG

Dr. Peter Fischer has served as junior professor at the University of Freiburg’s Department of Computer Science since October 2011. He studied at the Technical University of Munich and also began his dissertation there, com- pleting it – following a stint at the University of Heidel- berg – at ETH Zurich, Switzerland. Fischer focuses primarily on system-oriented work in data management, for instance with real-time analyses and the structure of information systems. His research on the analy- sis of social networks is currently being funded by Baden-Württemberg’s state junior professorship program. Photo: private Further Reading Cheng, J. / Adamic, L. / Dow, P. A. et al. (2014): Can cascades be predicted? In: Proceedings of the 23rd International Conference on World Wide Web, pp. 925–936. doi: 10.1145/2566486.2567997 Taxidou, I. / Fischer, P. M. (2014): Online analysis of information diffusion in Twitter. In: WWW '14 Companion Proceedings of the 23rd International Conference on World Wide Web, pp. 1313–1318. doi: 10.1145/2567948.2580050 Guille, A. / Hacid, H. / Favre, C. et al. (2013): Information diffusion in online social networks: A survey. In: SIGMOD Record 42/2, pp. 17–28. doi: 10.1145/2483574.2483575 A message spreads through the Internet: The illustration shows the reconstruction and visualization of an information cascade. The red circle represents the source. Source: Peter Fischer accuracy. The researchers can use the informa- tion to predict how the message will continue to spread in the next hours and days. What they don’t know is which topics will become a trend and which won’t. “We leave the content out of our analyses,” says Fischer. It would be much too time-consuming to ask each user individu- ally about their motives for sharing something. A second important aspect of Fischer’s research is the question as to the role of individual users. He analyzes the be- havior of Twitter users and attempts to identify which ones play a key role in the circulation of particular mes- sages. Which users act as disseminators, actively sharing content they find interesting? Who is primarily a passive con- sumer, reading information but not spreading it any further? Identifying such users could be interesting for com- panies. “Let’s say you want to advertise a new product and reach a million people with your message. You have a certain budget at your disposal – and it would of course be very helpful if you could reach precisely those people who find the message interesting and pass it on to others.” Ideally, the company would use an app to display in real time how the message spreads – and the original sender could influence who receives it and how. But developers are still a long way from being able to write such a program. What Peter Fischer is focusing on at the mo- ment is determining the connections between the dissemination patterns and the roles of the individual users. If he succeeds, the users might be able to better differentiate between trustworthy and untrustworthy information. It would also make it possible to expose spam more quickly. The predictions will not be definitive, however, because another factor that has thus far eluded all attempts at scientific analysis plays too large a role in the dissemination of messages: chance. 15uni wissen 01 2016 15uni wissen 012016