Detecting Aggression and Toxicity using a Multi Dimension Capsule Network

Abstract

In the era of social media, hate speech, trolling and verbal abuse have become a common issue. We present an approach to automatically classify such statements, using a new deep learning architecture. Our model comprises of a Multi Dimension Capsule Network that generates the representation of sentences which we use for classification. We further provide an analysis of our model's interpretation of such statements. We compare the results of our model with state-of-art classification algorithms and demonstrate our model's ability. It also has the capability to handle comments that are written in both Hindi and English, which are provided in the TRAC dataset. We also compare results on Kaggle's Toxic comment classification dataset.

Hierarchical Multi-label Classification of Text with Capsule Networks

“…They encapsulate features in different capsules and use the routing algorithm to cluster features for each task. Further applications to NLP span aggression, toxicity and emotion detection (Srivastava et al, 2018;Rathnayaka et al, 2018), embedding creation for knowledge graph completion (Nguyen et al, 2019), and knowledge transfer of user intents (Xia et al, 2018). Despite the suitable properties of capsule networks to classify into hierarchical structured categories, they have not yet been applied to HMC.…”

IMPACT OF THE SYNTACTIC DEPENDENCIES IN THE SENTENCES ON THE QUALITY OF THE IDENTIFICATION OF THE TOXIC COMMENTS IN THE SOCIAL NETWORKS (Ukr)

“…Наукові праці ВНТУ, 2019, № 4 2 При цьому виникають труднощі зі статистикою поганих слів. Погані слова автори токсичних коментарів навмисно спотворюють, наприклад, замість shit, пишуть shiiit, sh1t, sh!t, shi*, shyt, siht, тому науковці розроблюють спеціальні технології виявлення замаскованих образливих слів [3,4] Порядок слів враховують за деякою множиною усталених словосполучень, наприклад, за n-грамами [5], але це суттєво збільшує обчислювальну складність побудови моделей та не завжди розкриває семантику коментаря.…”

Detecting context abusiveness using hierarchical deep learning

“…They show that their method is able to capture both word sequence and order information in short texts compared to all the previous deep learning models. Srivastava et al (2019) pre-sented an approach that automatically classifies a toxic comment using a Multi Dimension Capsule Network. They also provide an analysis of their model's interpretation.…”