Georges Arsene Kamhoua
Florida International University
Georges Kamhoua is a Ph.D. candidate at Florida International University’s School of Computing and Information Sciences under the supervision of Dr. Niki Pissinou. He joined FIU in Fall 2014. He received his B.S. degree in Electrical Engineering and an M.S. degree in Electronics from University of Dschang, Cameroon in 2007 and 2012, respectively. His current research interests include security, online social networks, mobile wireless sensor networks, mobile computing, and crowdsourcing.
Online Social Networks (OSNs) and crowdsourcing services have rapidly grown into a wide network and offer users a variety of benefits such as sharing private information and gathering opinions. Users see OSNs not only as a platform to establish contacts but also as a means of soliciting contributions from a large group of people, exploiting the various applications existing in. In addition, crowdsourcing can be used for opinion gathering, rating, and the related weakness to ratings or reviews. Crowdsourcing marketplaces or internet crowdsourcing systems exploit the power of human workers available across the world to perform complex tasks to overcome the problems intrinsically difficult for computers to accomplish on their own. However, these services bring new threats to the community such as colluding Identity Clone Attack (ICA) and Non-Adversarial collusion respectively in OSNs and crowdsourcing. ICA uses the OSN infrastructure to endanger user privacy and security. The main purpose is to clone an existing user to infiltrate a certain social circle to gather information normally shared with the trusted users. In non-adversarial colluders, colluding workers choose one of them to do the work and the others purely copy and make minor changes to the original one to avoid being caught. Therefore, OSNs became a major weapon for launching cyberattacks on an organization and its people. In crowdsourcing one of the most important issues in using other people’s feedback to analyze the quality of tasks and the credibility of feedback (such as opinion gathering).
There is a great need for stronger defenses against cyber attacks through OSNs and crowdsourcing. This proposal addresses this need by focusing on a collaborative malicious group of users called collusion, which is more knowledgeable and difficult to deal with than an individual malicious user. We will develop new defense techniques to detect colluding groups in OSNs and overlapping colluding groups with different sizes in crowdsourcing. First, to resist colluding Identity Clone Attacks, we built a classifier based on supervised learning techniques. Second, we propose a semantic similarity measure and use community detection algorithms to overcome the behavior characteristics of non-adversarial colluders in crowdsourcing. Our preliminary results of these proposed methods indicate promising outcomes.