Keynote at ADCOM 2016
Title: Identifying (Some) Bad Actors in Social Media
Abstract: The Internet is full of bad actors including bots on Twitter, vandals on Wikipedia, and trolls on many different platforms. The repercussions of such acts have ranged from financial loss to loss of life. In this talk, I will discuss machine learning and engineering techniques to detect three kinds of bad actors. First, I will discuss the presence of bots on Twitter during a 10-month study of the 2014 Indian election and what we learned about characteristics distinguishing bot behavior from human behavior. I will also discuss the DARPA Twitter Bot challenge held in 2015 where the goal was to identify bots seeking to influence opinion on a very specific topic. Second, I will discuss the problem of identifying vandals on Wikipedia as such vandals can severely compromise the integrity of data in Wikipedia that millions of people worldwide rely on. Third, I will discuss techniques to identify trolls on social platforms such as Slashdot. If further time is available, I will discuss the use of data mining and machine learning techniques to identify other kinds of bad actors.
V.S. Subrahmanian is Professor of Computer Science and Director of the Lab for Computational Cultural Dynamics and Director of the Center for Digital International Government at the University of Maryland. His work stands squarely at the intersection of big data analytics for increased security, policy, and business needs.
Prof. Subrahmanian is one of the world leaders in logical reasoning with uncertainty, probabilistic logics, temporal probabilistic logics, and managing huge, heterogeneous databases with incomplete and inconsistent information, and multimedia databases. In recent years, he has developed scalable methods to apply probabilistic logic models to a wide variety of real-world scenarios. He created the field of computational cultural dynamics with a suite of novel methods to analyze the behaviors of terrorist groups and applied them to making forecasts and suggesting policies to shape behaviors of groups like Hezbollah, Lashkar-e-Taiba, and Indian Mujahideen. These methods have since been applied to a variety of problems including predicting the stability of nations, predicting when protests will turn violent, predicting systemic banking crises in countries, predicting spread of malware in countries, predicting health care outcomes and in manufacturing. In combination with social network analysis, these methods have also been used to identify bad actors on social media, forecast diffusion in social media, suggest methods to influence social networks. Prof. Subrahmanian led the team that won DARPA’s Twitter Influence Bot Detection Challenge under their SMISC program. Prof. Subrahmanian is one of the world leaders in the design, analysis, and application of big data analytics to real world problems so that optimal decisions can be made by governments and companies. In cyber-security, Prof. Subrahmanian developed some of the first secure query processing algorithms, flexible authentication frameworks, unexplained behavior detection and scalable detection of known threats. His Global Cyber-Vulnerability Report published in January, 2016, charracterizes cyber-risk of 44 countries by studying data on over 44 hosts per year over 2 years over 20 Billion telemetry and malware reports.
In addition to his academic work, Prof. Subrahmanian has served for 2 years (2013, 2014) on the US-India Strategic Dialog (track 2), the India-Israel track 2/dialog (2013, 2014), the US Air Force Science Advisory Board, and DARPA”s Executive Advisory Council on their Advanced Logistics Program. He serves currently on the Research Advisory Board of Tata Consultancy Services (India’s biggest software firm), the Board of Directors of the Development Gateway (created by the World Bank in 1999), and Sentimetrix, Inc., a big data analytics firm, and CosmosId, a leading bioinformatics company. He serves on the boards of numerous journals including Science, ACM Transactions on Intelligent Systems & Technology, ACM Transactions on Computational Logic, and IEEE Transactions on Computational Social Systems.