Detection of novel social bots by ensembles of specialized classifiers

Sayyadiharikandeh, Mohsen and Varol, Onur and Yang, Kai-Cheng and Flammini, Alessandro and Menczer, Filippo (2020) Detection of novel social bots by ensembles of specialized classifiers. In: CIKM '20: Proceedings of the 29th ACM International Conference on Information & Knowledge Management, Ireland

[thumbnail of 3340531.3412698.pdf] PDF

Download (3MB)


Malicious actors create inauthentic social media accounts controlled in part by algorithms, known as social bots, to disseminate misin- formation and agitate online discussion. While researchers have developed sophisticated methods to detect abuse, novel bots with diverse behaviors evade detection. We show that different types of bots are characterized by different behavioral features. As a result, supervised learning techniques suffer severe performance deterioration when attempting to detect behaviors not observed in the training data. Moreover, tuning these models to recognize novel bots requires retraining with a significant amount of new annotations, which are expensive to obtain. To address these issues, we propose a new supervised learning method that trains classi- fiers specialized for each class of bots and combines their decisions through the maximum rule. The ensemble of specialized classifiers (ESC) can better generalize, leading to an average improvement of 56% in F1 score for unseen accounts across datasets. Furthermore, novel bot behaviors are learned with fewer labeled examples during retraining. We deployed ESC in the newest version of Botometer, a popular tool to detect social bots in the wild, with a cross-validation AUC of 0.99.
Item Type: Papers in Conference Proceedings
Uncontrolled Keywords: social media, social bots, machine learning, cross-domain, recall
Divisions: Faculty of Engineering and Natural Sciences
Depositing User: Onur Varol
Date Deposited: 29 Aug 2021 16:46
Last Modified: 26 Apr 2022 09:38

Actions (login required)

View Item
View Item