Benchmarking Aggression Identification in Social Media

By Ritesh Kumar, Atul Kr. Ojha, Shervin Malmasi, Marcos Zampieri. 2018

447 citations recorded by the source corpus.

Subfields: Natural Language Processing, Evaluation & Benchmarks, 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 Natural Language Processing

In Evaluation & Benchmarks

In Safety, Robustness & Interpretability

Citations in the corpus

The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes (2020, background): “Several text-only hate speech datasets have been released, mostly based on Twitter [83, 82, 11, 24, 20], and various architectures have been proposed for classifiers [45, 54, 53].”
Build it Break it Fix it for Dialogue Safety: Robustness from Adversarial Human Attack (2019, methodology): “The best performing systems in each of the competitions mentioned above (for aggression and toxic comment classiﬁcation) used deep learning approaches such as LSTMs and CNNs (Kumar et al., 2018; Google, 2018).”
Learning from the Worst: Dynamically Generated Datasets to Improve Online Hate Detection (2021, background): “It is similar to the ‘implicit’ and ‘covert’ categories used in other taxonomies (Waseem et al., 2017; Vidgen and Yasseri, 2019; Kumar et al., 2018).”
Challenges and frontiers in abusive content detection (2019, neutral): “Elsewhere, Kumar et al. distinguish ‘overt’ from ‘covert’ hate (Kumar, Ojha, Malmasi, & Zampieri, 2018).”
SemEval-2019 Task 6: Identifying and Categorizing Offensive Language in Social Media (OffensEval) (2019, methodology): “, 2011), and aggression (Kumar et al., 2018).”
Predicting the Type and Target of Offensive Posts in Social Media (2019, methodology): “, 2017), (cyber-)aggression (Kumar et al., 2018), (cyber-)bullying (Xu et al.”
Detecting and Classifying Malevolent Dialogue Responses: Taxonomy, Data and Methodology (2020, background): “Kumar et al. [30] use the extent of aggression as classification categories in their dataset Trolling, Aggression and Cyberbullying (TRAC): ‘overtly aggressive’, ‘covertly aggressive’ and ‘non-aggressive’.”
SemEval-2020 Task 12: Multilingual Offensive Language Identification in Social Media (OffensEval 2020) (2020, background): “Some teams used additional datasets from HASOC 2019 (Mandl et al., 2019), the Kaggle competitions on Detecting Insults in Social Commentary 6 and Toxic Comment Classiﬁcation 7 , TRAC 2018 shared task on Aggression Identiﬁ-cation (Kumar et al., 2018a; Kumar et al., 2018c), Wikipedia Detox dataset (Wu”
Resources and benchmark corpora for hate speech detection: a systematic review (2020, background): “…with retrieving the whole timeline from users or pages considered hateful, i.e., where it is likely to ﬁnd hateful contents (Mubarak et al. 2017; Kumar et al. 2018a), or from discussion threads about controversial topics that can easily trigger a certain language (Hammer 2017), taking into…”
Just Say No: Analyzing the Stance of Neural Dialogue Generation in Offensive Contexts (2021, background): “…social media posts or comments while ignoring the context (Davidson et al., 2017; Xu et al., 2012; Zampieri et al., 2019; Rosenthal et al., 2020; Kumar et al., 2018; Garibo i Orts, 2019; Ousidhoum et al., 2019; Breitfeller et al., 2019; Sap et al., 2020; Hada et al., 2021; Barikeri et al.,…”
Tackling Online Abuse: A Survey of Automated Abuse Detection Methods (2019, background): “Kumar et al. (2018) noted similar trends in the shared task on aggression identiﬁcation on DATA - FACEBOOK .”
Towards multidomain and multilingual abusive language detection: a survey (2021, methodology): “Abusive language is generally used as an umbrella term [143], covering several sub-categories, such as cyberbullying [55, 131], hate speech [33, 144], toxic comments [150], offensive language [152] and online aggression [67].”
SOLID: A Large-Scale Semi-Supervised Dataset for Offensive Language Identification (2020, methodology): “There were two editions of the TRAC shared task on Aggression Identiﬁcation (Kumar et al., 2018, 2020), which provided participants with datasets containing annotated Facebook posts and comments in English and Hindi for training and validation.”
Studying Generalisability across Abusive Language Detection Datasets (2019, background): “Datasets for these tasks have been collected from various social media platforms, such as Twitter (Waseem and Hovy, 2016; Davidson et al., 2017; Founta et al., 2018; Burnap and Williams, 2015; Golbeck et al., 2017), Facebook (Kumar et al., 2018a), Instagram (Hosseinmardi et al., 2015; Zhong et al., ”
Abusive Language Detection using Syntactic Dependency Graphs (2020, background): “Further, Kumar et al. (2018) concluded that with optimal feature selection, classiﬁers like SVM and Random forest performs at par with neural networks.”
Transformer models for text-based emotion detection: a review of BERT-based approaches (2021, methodology): “With data from the shared task on aggression identification organized at the Trolling, Aggression, and Cyberbullying (TRAC-1) work-shop (Kumar et al. 2018), texts were classified into three categories: overtly aggressive (OAG), covertly aggressive (CAG) or non-aggressive (NAG).”