Leveraging Intra-User and Inter-User Representation Learning for Automated Hate Speech Detection

By Jing Qian, Mai Elsherief, Elizabeth M. Belding-Royer, William Yang Wang. 2018

93 citations recorded by the source corpus.

Hate speech detection is a critical, yet challenging problem in Natural Language Processing (NLP). Despite the existence of numerous studies dedicated to the development of NLP hate speech detection approaches, the accuracy is still poor. The central problem is that social media posts are short and noisy, and most existing hate speech detection solutions take each post as an isolated input instance, which is likely to yield high false positive and negative rates. In this paper, we radically improve automated hate speech detection by presenting a novel model that leverages intra-user and inter-user representation learning for robust hate speech detection on Twitter. In addition to the target Tweet, we collect and analyze the user’s historical posts to model intra-user Tweet representations. To suppress the noise in a single Tweet, we also model the similar Tweets posted by all other users with reinforced inter-user representation learning techniques. Experimentally, we show that leveraging these two representations can significantly improve the f-score of a strong bidirectional LSTM baseline model by 10.1%.

Subfields: Artificial Intelligence, Information Systems, Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

HateCheck: Functional Tests for Hate Speech Detection Models (2020, background): “The use of humour, for instance, has been highlighted as an important challenge for hate speech research (van Aken et al., 2018; Qian et al., 2018; Vidgen et al., 2020a).”
Challenges for Toxic Comment Classification: An In-Depth Error Analysis (2018, methodology): “We analyse common error classes of our ensemble based on research from Zhang and Luo (2018); Zhang et al. (2018); Qian et al. (2018); Davidson et al. (2017); Schmidt and Wiegand (2017); No-bata et al. (2016).”
HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection (2020, methodology): “To tackle this, various methods have been proposed for hate speech detection (Burnap and Williams 2016; Ribeiro et al. 2018; Zhang, Robinson, and Tepper 2018; Qian et al. 2018a).”
Returning the N to NLP: Towards Contextually Personalized Classification Models (2020, background): “, 2018), abusive language use (Qian et al., 2018; Mishra et al., 2018) or suicide ideation (Mishra et al.”
Predicting Anti-Asian Hateful Users on Twitter during COVID-19 (2021, background): “…in online discussions, Ribeiro et al. (2018) detect the users who express online hate by checking their profiles or linguistic characteristics, Qian et al. (2018) incorporate users’ historical tweets and user similarities for hate speech detection, and Lyu et al. (2020) compares hate…”
Handling Bias in Toxic Speech Detection: A Survey (2022, background): “This high author bias can be attributed to the improvements in performance when encoding the user meta-information as features [63, 75].”
Lifelong Learning of Hate Speech Classification on Social Media (2021, methodology): “Therefore, instead of differentiating normal posts from the other offensive ones, Qian et al. (2018a) propose a more ﬁne-grained hate speech classiﬁcation task that attributes hate groups to individual tweets.”
Developing an online hate classifier for multiple social media platforms (2020, methodology): “BERT has vastly outper-formed previous models, such as the Generative Pretrained Transformer (GPT) [63] and Embeddings from Language Models (ELMo) [24], in tasks such as question answering [92], named-entity recognition [1], and natural language inference [76], among others.”
Aggression Detection in Social Media: Using Deep Neural Networks, Data Augmentation, and Pseudo Labeling (2018, background): “Rather than rely only on the textual content of social media posts in identifying hate speech, (Qian et al., 2018; Founta et al., 2018) investigated the use of metadata about the users or the posts.”
Tackling Online Abuse: A Survey of Automated Abuse Detection Methods (2019, methodology): “Qian et al. (2018) used LSTM s for modeling inter and intra-user relationships on DATA - TWITTER - WH , with sexist and racist tweets combined into one category.”
Detect All Abuse! Toward Universal Abusive Language Detection Models (2020, methodology): “Many social media companies and researchers have utilised multiple resources, including machine learning, human reviewers and lexicon-based text analytics to detect abusive language (Waseem, 2016; Qian et al., 2018).”
User Preference-aware Fake News Detection (2021, background): “Similar to [2, 13, 20] which model the users’ personality, sentiment and stance using their historical posts, we leverage the historical posts of a user to encode his/her preference implicitly.”
Exploring the Role of Grammar and Word Choice in Bias Toward African American English (AAE) in Hate Speech Classification (2022, methodology): “These and other deep learning models for this task mostly rely on a few popular approaches including Convolutional Neural Nets [1], Recurrent Neural Nets [30], Bi-LSTM [33], and word embeddings (ie BERT, Word2Vec, and TF-IDF Vectorizor), many of which combine these techniques [1].”
Misogyny Detection in Twitter: a Multilingual and Cross-Domain Study (2020, unknown): “…hate speech detection Badjatiya, Gupta, Gupta, and Varma (2017); Fehn Unsvåg and Gambäck (2018); Kshirsagar, Cukuvac, McKeown, and McGregor (2018); Qian, ElSherief, Belding, and Wang (2018), author profiling in abuse detection Mishra, Tredici, Yannakoudakis, and Shutova (2018), bias in abusive…”