Fraud detection in Crowdsourcing platforms
Crowdsourcing is the art of involving individuals in the Internet for the resolution of industrial tasks. Apart from the intrinsic quality and management of a large volume of tasks and workers that current crowdsourcing platforms must face daily, crowdsourcing is very sensible towards new threat scenarios, or fraudulent behaviour, executed by the individuals participating in the platform. For example, several quality assurance mechanisms are based in the assumption that evaluation tasks can also be performed by other individuals in the platform. Even though this procedure is efficient when applied in traditional organizations, individuals in the crowd could self-organize to achieve both the task executors and reviewers at the same time.
This aspect of crowdsourcing is very relevant as industry is mainly driving motivation of individuals through financial rewards. During this project, we aim at tackling this challenge by (1) studying and formalizing the application of crowdsourcing technologies in the industry, (2) enhance state of the art quality assurance mechanisms and (3) working on the active detection of fraudulent behaviour.