Understanding Abuse: A Typology of Abusive Language Detection Subtasks

Waseem, Zeerak; Davidson, Thomas; Warmsley, Dana; Weber, Ingmar

doi:10.48550/arxiv.1705.09899

Cited by 27 publications

(29 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NLP-aided approaches to detect abusive behavior online is an active research area (Schmidt and Wiegand, 2017;Mishra et al, 2019;Corazza et al, 2020). Researchers have developed typologies of online abuse (Waseem et al, 2017), constructed datasets annotated with different types of abusive language (Warner and Hirschberg, 2012;Price et al, 2020;Vidgen et al, 2021), and built NLP models to detect them efficiently Mozafari et al, 2019). Researchers have also expanded the focus to more subtle forms of abuse such as condescension and microaggressions (Breitfeller et al, 2019;.…”

Section: Detecting Online Abusementioning

confidence: 99%

Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations

Davani¹,

Díaz²,

Prabhakaran³

2021

Preprint

View full text Add to dashboard Cite

Majority voting and averaging are common approaches employed to resolve annotator disagreements and derive single ground truth labels from multiple annotations. However, annotators may systematically disagree with one another, often reflecting their individual biases and values, especially in the case of subjective tasks such as detecting affect, aggression, and hate speech. Annotator disagreements may capture important nuances in such tasks that are often ignored while aggregating annotations to a single ground truth. In order to address this, we investigate the efficacy of multi-annotator models. In particular, our multi-task based approach treats predicting each annotators' judgements as separate subtasks, while sharing a common learned representation of the task. We show that this approach yields same or better performance than aggregating labels in the data prior to training across seven different binary classification tasks. Our approach also provides a way to estimate uncertainty in predictions, which we demonstrate better correlate with annotation disagreements than traditional methods. Being able to model uncertainty is especially useful in deployment scenarios where knowing when not to make a prediction is important.

show abstract

Section: Detecting Online Abusementioning

confidence: 99%

Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations

Davani¹,

Díaz²,

Prabhakaran³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Prior work has also explored sophisticated feature representations including n-grams, linguistic and syntactic features (Nobata et al, 2016), TF-IDF (Salminen et al, 2018), Bag of Words and word embeddings (Djuric et al, 2015), as well as content-specific features such as mentions, proper nouns, named entities, and target group specific vocabularies (Waseem et al, 2017). Topic modeling approaches such as Labeled Latent Dirichlet Allocation have also been proposed (Saleem et al, 2017).…”

Section: Related Workmentioning

confidence: 99%

Few-shot Instruction Prompts for Pretrained Language Models to Detect Social Biases

Prabhumoye¹,

Kocielnik²,

Shoeybi³

et al. 2021

Preprint

View full text Add to dashboard Cite

Warning: this paper contains content that may be offensive or upsetting.Detecting social bias in text is challenging due to nuance, subjectivity, and difficulty in obtaining good quality labeled datasets at scale, especially given the evolving nature of social biases and society. To address these challenges, we propose a few-shot instructionbased method for prompting pre-trained language models (LMs). We select a few labelbalanced exemplars from a small support repository that are closest to the query to be labeled in the embedding space. We then provide the LM with instruction that consists of this subset of labeled exemplars, the query text to be classified, a definition of bias, and prompt it to make a decision. We demonstrate that large LMs used in a few-shot context can detect different types of fine-grained biases with similar and sometimes superior accuracy to fine-tuned models. We observe that the largest 530B parameter model is significantly more effective in detecting social bias compared to smaller models (achieving at least 20% improvement in AUC metric compared to other models). It also maintains a high AUC (dropping less than 5%) in a few-shot setting with a labeled repository reduced to as few as 100 samples. Large pretrained language models thus make it easier and quicker to build new bias detectors.

show abstract

“…Sexist or racial slur, attacking a minority, seeking to silence a minority, criticizing a minority or defending xenophobia are some of their criterias. In similar work, Waseem et al [14] studied and provided an assessment of influence of annotator knowledge on hate specch on twitter. In other work Wijesiriwardene et al [16] showed that individual tweets are not sufficient to provide evidence for toxic behaviour instead context in interactions can give a better explanation.…”

Section: Related Workmentioning

confidence: 99%

Divide and Conquer: An Ensemble Approach for Hostile Post Detection in Hindi

Bhatnagar

Kumar

Moghili

et al. 2021

Combating Online Hostile Posts in Regional Languages During Emergency Situation

View full text Add to dashboard Cite

Recently the NLP community has started showing interest towards the challenging task of Hostile Post Detection. This paper presents our system for Shared Task @ Constraint2021 on "Hostile Post Detection in Hindi" 1 . The data for this shared task is provided in Hindi Devanagari script which was collected from Twitter and Facebook. It is a multi-label multi-class classification problem where each data instance is annotated into one or more of the five classes: fake, hate, offensive, defamation, and non-hostile. We propose a two level architecture which is made up of BERT based classifiers and statistical classifiers to solve this problem. Our team 'Albatross', scored 0.9709 Coarse grained hostility F1 score measure on Hostile Post Detection in Hindi subtask and secured 2nd rank out of 45 teams for the task 2 . Our submission is ranked 2nd and 3rd out of a total of 156 submissions with Coarse grained hostility F1 score of 0.9709 and 0.9703 respectively. Our fine grained scores are also very encouraging and can be improved with further finetuning. The code is publicly available 3 .

show abstract

Understanding Abuse: A Typology of Abusive Language Detection Subtasks

Cited by 27 publications

References 0 publications

Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations

Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations

Few-shot Instruction Prompts for Pretrained Language Models to Detect Social Biases

Divide and Conquer: An Ensemble Approach for Hostile Post Detection in Hindi

Contact Info

Product

Resources

About