A Deep Learning Approach for Robust Detection of Bots in Twitter Using Transformers

During the last decades, volume of multimedia content posted in social networks has grown exponentially and such information is immediately propagated consumed by a significant number users. In this scenario, disruption fake news providers bot accounts for spreading propaganda as well sensitive throughout network fostered applied research to automatically measure reliability via Artificial Intelligence (AI). paper, we present multilingual approach addressing identification task Twitter Deep learning (DL) approaches support end-users when checking credibility certain account. To do so, several experiments were conducted using state-of-the-art Multilingual Language Models generate an encoding text-based features user account that are later on concatenated with rest metadata build potential input vector top Dense Network denoted Bot-DenseNet. Consequently, paper assesses language constraint from previous studies where only considered either or together some basic semantic text features. Moreover, Bot-DenseNet produces low-dimensional representation which can be used any application within Information Retrieval (IR) framework.