Tutorial Abstract

As social media continues to drive the discussion of large scale events, the presence of malicious users grows with it. This year, Twitter reports that over half of the accounts created have been deleted due to malicious activity. Many of these malicious users are "bots", non-human users controlled by a computer program. In this tutorial, we focus on malicious bots: those that focus on swaying the general discussion on Twitter, usually for political purposes. This tutorial will focus on recent approaches to the problem of finding and mitigating these bots. We will review state-of-the-art methodologies in the social network analysis, data mining, and machine learning literature that have been proposed to detect bots, their controllers, and to intercept these bots to allow the voice of real users on social media to be heard. We will also review the next wave of challenging problems that come with investigating this important problem on social media.

Tutorial Outline

Section Duration
Introduction 15 minutes
Network Methods 40 minutes
Groud Truth Discovery 20 minutes
Machine Learning Techniques 20 minutes
Conclusion 15 minutes

Tutorial Slides

Tutorial Slides

Reference Materials

Data Code Further Reading

Presenters

Page last updated: April 14, 2016.