Dr. Nitin Agarwal, the Maulden-Entergy Chair and Distinguished Professor of Information Science and the founding director of COSMOS Research Center, has received a $95,227 grant from the U.S. National Science Foundation (NSF) flow-through from Arkansas Economic Development Commission’s (AEDC) Division of Science and Technology. Dr. Agarwal said, “the grant will help develop novel big data mining algorithms for a safer social media.”
Online Social Networks (OSNs) have revolutionized how societies interact. While this new phenomenon in online socialization has brought the world closer, OSNs have also led to new vectors to facilitate cybercrime, cyberterrorism, cyberwarfare, and other deviant behaviors perpetrated by state/non-state actors.
Due to afforded anonymity and perceived less personal risk of connecting and acting online, deviant groups are becoming increasingly common among socio-technically competent ‘hacktivist’ groups to provoke hysteria, spread mis/disinformation, coordinate (cyber)-attacks, or even effect civil conflicts. Such deviant behaviors are categorized as the new face of transnational crime organizations (TCOs) that could pose significant risks to social, political, and economic stability.
Online deviant groups have grown in parallel with OSNs, whether it is black hat hackers using Twitter to recruit and arm attackers, announce operational details, coordinate cyberattacks, and post instructional or recruitment videos on YouTube targeting certain demographics; or state/non-state actors’ and extremist groups’ savvy use of social communication platforms to conduct phishing operations, such as viral retweeting a message containing image which if clicked unleashes malware. The threat these deviant groups pose is real and can manifest in several forms of deviance, such as the disabling of critical infrastructure (e.g., the Ukraine power outage caused by Russian-sponsored hackers that coordinated a cyberattack in December 2015).
Since OSNs, hosted on platforms like Twitter, Facebook, YouTube, and blogs, are continuously producing data with tremendous volume, variety, and velocity – big data at its essence, traditional methods of forensic investigation would be insufficient, as this data would be real-time, constantly expanding, and simply not found in traditional sources of forensic evidence. These newer forms of data, especially the communications of hacker groups on OSNs, would offer insights to coordination and planning, for example. Social media is growing as a data source for cyber forensics, providing new types of artifacts that can be relevant to investigations. Practitioners must embrace the idea of using real-time intelligence to assist in cyber forensic investigations, and not just post-mortem data.
Dr. Agarwal said, “we need to expand the traditional definitions of cyber threats from hardware attacks and malware infections to include such insidious threats that influence behaviors and actions.” Observable malicious behaviors in OSNs, similar to the aforementioned ones, continue to negatively impact society warranting their scientific inquiry.
To stem access of malicious actors and limit their influence campaigns, decision makers often resort to limit access to these platforms and sometimes even shut down the internet. These methods affect both benevolent and malicious campaigns alike. There is a need for efficient methods that detect coordinating groups of malicious users to dismantle their influence campaign networks.
Dr. Agarwal said, “main objective of this research is to identify and track malicious groups in real-world networks that hide in plain sight. This research proposes a way to identify key groups, investigate their interest, and predict their influence in the network.”
Conventional methods have focused on identifying community structures in OSNs and are oblivious to these key coordinating groups and the various contexts in which they share interests. Moreover, as OSNs grow, groups reorganize, and the network structural properties change, making it difficult to identify these key groups and their information diffusion networks.
The outcomes of the research will be published in scientific journals and also as a publicly available web-based tool that will complement the suite of analytical tools developed by COSMOS, viz., BlogTracker, VTracker, and COVID-19 Misinformation Tracker.
Dr. Agarwal said, “the grant will further create training opportunities for our students to build a data science workforce that can develop technically and ethically competent solutions to real-world problems.”