We are presenting at EMNLP in Brussels

We will be presenting at the Conference on Empirical Methods in Natural Language Processing (EMNLP) on Abusive Language Online (ALW2), organized by the Association for Computational Linguistics (ACL) special interest group on linguistic data and corpus-based approaches to NLP (SIGDAT). Our presentation will focus on how online communication through various means such as comment threads, profile pages, social media can be made better by reducing negative outcomes such as cyberbullying, hate speech, scapegoating, etc.

Our goal is to bring together NLP researchers with victims of abusive language, free speech advocates, sociologists, and legal experts to advance the knowledge and provide insights from our work.

The paper, “A Review of Standard Text Classification Practices for Multi-label Toxicity Identification of Online Content”, written by Isar Nejadgholi and Isuru Gunasekara will be discussed during this presentation and include demonstrations in the form of a live demo. This work focuses on a classification system that can identify sentences which are toxic, severe toxic, obscene, threat, insult, or identity hate. In this work, we explore different text classification methods including classical NLP methods such as TFIDF in addition to presenting two deep neural network (DNN) models and a gradient boosting method to combine the results of two DNN classifiers.

Henri Kuschkowitzconference