Learning to Detect Online Harassment on Twitter with the Transformer

Bugueño, Margarita Constanza; Mendoza, Marcelo

doi:10.1007/978-3-030-43887-6_23

Cited by 16 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Harassment. Many early researchers defined harassment as a binary classification -a social media post (e.g., a tweet) is either harassing or not [37,38,39,40].…”

Section: Need For Empirical Modelsmentioning

confidence: 99%

Defining and Detecting Toxicity on Social Media: Context and Knowledge are Key

Sheth¹,

Shalin²,

Kursuncu³

2021

Preprint

View full text Add to dashboard Cite

As the role of online platforms has become increasingly prominent for communication, toxic behaviors, such as cyberbullying and harassment, have been rampant in the last decade. On the other hand, online toxicity is multi-dimensional and sensitive in nature, which makes its detection challenging. As the impact of exposure to online toxicity can lead to serious implications for individuals and communities, reliable models and algorithms are required for detecting and understanding such communications. In this paper, we define toxicity to provide a foundation drawing social theories. Then, we provide an approach that identifies multiple dimensions of toxicity and incorporates explicit knowledge in a statistical learning algorithm to resolve ambiguity across such dimensions.

show abstract

“…Harassment. Many early researchers defined harassment as a binary classification -a social media post (e.g., a tweet) is either harassing or not [37,38,39,40].…”

Section: Need For Empirical Modelsmentioning

confidence: 99%

Defining and Detecting Toxicity on Social Media: Context and Knowledge are Key

Sheth¹,

Shalin²,

Kursuncu³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Park and Fung (2017) designed a HybridCNN (word-level and character-level) model on abusive tweet detection in both one-step and two-step style. Several works have applied bidirectional Gated Recurrent Unit (Bi-GRU) networks with Latent Topic Clustering (LTC) Lee et al (2018) and a a transformer-based framework Bugueño and Mendoza (2019). Some works integrated user profiling into their ALD models.…”

Section: Ald Approachesmentioning

confidence: 99%

“…Three Bi-GRU layers with hidden dimensions 128, 128, and 64 are stacked, and the model is trained using learning rate 4e-5 with 10 epochs. Character-based Transformer (CBT): CBT (Bugueño and Mendoza, 2019) uses a transformer-based classifier with Chars2vec embeddings. Transformer encoders with hidden dimension 400, learning rate 4e-6 with 3 epochs are used.…”

Section: Evaluation Settingmentioning

confidence: 99%

Detect All Abuse! Toward Universal Abusive Language Detection Models

Wang

Han³

et al. 2020

Proceedings of the 28th International Conference on Computational Linguistics

View full text Add to dashboard Cite

Online abusive language detection (ALD) has become a societal issue of increasing importance in recent years. Several previous works in online ALD focused on solving a single abusive language problem in a single domain, like Twitter, and have not been successfully transferable to the general ALD task or domain. In this paper, we introduce a new generic ALD framework, MACAS, which is capable of addressing several types of ALD tasks across different domains. Our generic framework covers multi-aspect abusive language embeddings that represent the target and content aspects of abusive language and applies a textual graph embedding that analyses the user's linguistic behaviour. Then, we propose and use the cross-attention gate flow mechanism to embrace multiple aspects of abusive language. Quantitative and qualitative evaluation results show that our ALD algorithm rivals or exceeds the six state-of-the-art ALD algorithms across seven ALD datasets covering multiple aspects of abusive language and different online community domains. The code can be downloaded from https://github.com/usydnlp/MACAS.

show abstract

“…In the work of this project, early detection of users' stances, harassment detection, early fake news detection, and the study of polarization dynamics in social media occupy central roles. 17,18 Development of robust information structures. This project focuses on two prominent information disorders; that is, problems that work against the development of robust information structures.…”

Section: Imfd Aims To Be a Reference Center In Latin America Related To State-ofthe-art Research On The Foundational Problems With Data Amentioning

confidence: 99%

Chile's new interdisciplinary institute for foundational research on data

Arenas

Barceló

2020

Commun. ACM

View full text Add to dashboard Cite

Learning to Detect Online Harassment on Twitter with the Transformer

Cited by 16 publications

References 15 publications

Defining and Detecting Toxicity on Social Media: Context and Knowledge are Key

Defining and Detecting Toxicity on Social Media: Context and Knowledge are Key

Detect All Abuse! Toward Universal Abusive Language Detection Models

Chile's new interdisciplinary institute for foundational research on data

Contact Info

Product

Resources

About