- Truth Learning in a Social Setting
- Project Year:
2023
- REU Student (s):
Jordan Chong | New York University NY
| Matt Lu | Washington University in St Louis MO
- Student 1 Institution:
New York University
- Student 2 Institution:
Washington University in St Louis
- Project Mentor:
Jie Gao
- Project Mentor Area:
Computer Science
- Project Abstract:
We are interested in designing social networks which support the spreading of true information while limiting the ability for disinformation to spread. Having access more information doesn't always lead to discovering the truth as phenomena known as "information cascades" can lead to herd mentality propagating false information. Representing social networks with graphs, we studied two areas: high degree and low degree graphs. High degree graphs have the ability to quickly aggregate information and decipher truth into a few high value nodes with many neighbors. One example we studied was the preferential attachment model by Albert-Barabasi. In low or constant degree graphs, a more nuanced approach is needed to allow truth learning to succeed. We studied two examples of these: an adjusted Connected Caveman graph and a grid structure. Moreover, in large-scale networks such as news portals or content-sharing platforms, having a solid grasp of the underlying network structure and ordering helps optimize the process of dissemination and improve the accuracy of information spreading. In modern society, misinformation is a pervasive issue that can have large societal consequences. As we strive to combat the spread of false information, studying the ways information spreads in a variety of networks becomes crucial.
