Machine learning in social media: topic modeling, community detection and causal inference
Federico Albanese - University of Buenos Aires
Wednesday, 28 September, 2022 - 14:00
Aula B203, Secondo Piano - DIAG Sapienza - Via Ariosto 205
Even though the Internet and social media have increased the amount of news and information people can consume, most users are only exposed to content that reinforces their positions and isolates them from other ideological communities. This environment has real consequences with great impact on our lives like severe political polarization, easy spread of fake news, political extremism, hate groups and the lack of enriching debates, among others. Therefore, encouraging conversations between different groups of users and with different points of views is important for healthy societies.
In this talk, we will focus on how we can use machine learning models with millions of posts from Twitter and Reddit to characterize how users talk to each other. First, we will discuss how to apply popular topic modeling algorithms to tweets, which tend to be shorter and more incoherente than other text corpus, not only using the text of the tweet, but also the underlying interaction graph. We will show how this methodology could be applied to electoral context datasets in order to characterize the interests of users from different political leanings and, in particular, to community-changing users (people that change from one political community to another one during the campaign).
Finally, we will analyze the news-sharing behavior of users in Reddit and measure the causal impact that sharing an article with an opposing political leaning has on the toxicity of the online conversation.
Federico Albanese is a member of the Institute of calculus and PhD student in Computer Science at the University of Buenos Aires (UBA). His research focuses on machine learning applications on social media. He received his master degree in physics from the same university. Previously, he has worked as a data scientist at Hexagon doing financial analysis and did two internships at Facebook (now Meta) where he worked on deep bayesian networks and ranking models. He is currently an assistant professor in the master's degree in data mining at the UBA and advisor to two master's students on causal inference of social media messages.
ID riunione: 828 6166 0759