DBIS

Unfortunately, the online ecosystem is constantly under threat from malicious social bots that can spread worms, phishing links or spam, and manipulate user behavior on a particular social network. For example, there was a social bot used in Syria that flooded Twitter with hashtags related to the Syrian civil war with irrelevant topics that redirected users’ attention away from the war. In addition, social bots have recently been held responsible for falsifying online discussions about major political elections in Western countries, including the 2016 USA presidential election and the Brexit referendum in the UK. Twitter, for example, has been particularly hard hit, as bots make up a shockingly large number of its users. Detecting and removing malicious social bots on online social networks is therefore of utmost importance.

Artificial immune systems (AIS) use concepts and algorithms inspired by human immune system theory. In recent years, AIS have attracted significant interest due to their inherent robustness and ability to adapt to changes in the network as well as to complex, unknown, and dynamically changing environments. An overview of AIS algorithms can be found here.

AIS is able to adapt to the changing nature and behavior patterns of malicious social bots. Since malicious bots are always prone to make subtle changes and develop new ways to spread warm, phishing links and spam as widely as possible, this is a critical advantage. Algorithms that fail to detect these subtle changes will eventually fail to handle even slightly modified bot behavior.

The goal of this thesis is to develop a AIS for detection of malicious social bots in online social network.