BotSSCL: Social Bot Detection with Self-Supervised Contrastive Learning

TL;DR

BotSSCL introduces a self-supervised contrastive learning framework that significantly improves social bot detection accuracy, robustness, and generalizability across datasets and against adversarial manipulations.

Contribution

This paper presents a novel self-supervised contrastive learning approach for social bot detection, enhancing robustness and cross-dataset generalization over existing methods.

Findings

Outperforms baseline methods with 6-8% higher F1 scores.

Achieves 67% F1 when trained and tested on different datasets.

Reduces adversarial success rate to 4%.

Abstract

The detection of automated accounts, also known as "social bots", has been an increasingly important concern for online social networks (OSNs). While several methods have been proposed for detecting social bots, significant research gaps remain. First, current models exhibit limitations in detecting sophisticated bots that aim to mimic genuine OSN users. Second, these methods often rely on simplistic profile features, which are susceptible to manipulation. In addition to their vulnerability to adversarial manipulations, these models lack generalizability, resulting in subpar performance when trained on one dataset and tested on another. To address these challenges, we propose a novel framework for social Bot detection with Self-Supervised Contrastive Learning (BotSSCL). Our framework leverages contrastive learning to distinguish between social bots and humans in the embedding space to improve linear separability. The high-level representations derived by BotSSCL enhance its resilience to variations in data distribution and ensure generalizability. We evaluate BotSSCL's robustness against adversarial attempts to manipulate bot accounts to evade detection. Experiments on two datasets featuring sophisticated bots demonstrate that BotSSCL outperforms other supervised, unsupervised, and self-supervised baseline methods. We achieve approx. 6% and approx. 8% higher (F1) performance than SOTA on both datasets. In addition, BotSSCL also achieves 67% F1 when trained on one dataset and tested with another, demonstrating its generalizability. Lastly, BotSSCL increases adversarial complexity and only allows 4% success to the adversary in evading detection.