Sexism Identification in Tweets and Gabs using Deep Neural Networks

TL;DR

This paper evaluates deep neural network models, including LSTMs, CNNs, and transformer-based models like BERT, for automatic sexism detection in social media texts, achieving competitive results and highlighting challenges in subjectivity and language complexity.

Contribution

It compares various neural architectures and transfer learning techniques for sexism classification, demonstrating the effectiveness of BERT and CNNs with data augmentation.

Findings

BERT and multi-filter CNN perform best among models.

Data augmentation improves multi-class classification accuracy.

Model errors reveal challenges due to language complexity and subjectivity.

Abstract

Through anonymisation and accessibility, social media platforms have facilitated the proliferation of hate speech, prompting increased research in developing automatic methods to identify these texts. This paper explores the classification of sexism in text using a variety of deep neural network model architectures such as Long-Short-Term Memory (LSTMs) and Convolutional Neural Networks (CNNs). These networks are used in conjunction with transfer learning in the form of Bidirectional Encoder Representations from Transformers (BERT) and DistilBERT models, along with data augmentation, to perform binary and multiclass sexism classification on the dataset of tweets and gabs from the sEXism Identification in Social neTworks (EXIST) task in IberLEF 2021. The models are seen to perform comparatively to those from the competition, with the best performances seen using BERT and a multi-filter CNN model. Data augmentation further improves these results for the multi-class classification task. This paper also explores the errors made by the models and discusses the difficulty in automatically classifying sexism due to the subjectivity of the labels and the complexity of natural language used in social media.