On Transfer Learning for Detecting Abusive Language Online

Uban, Ana Sabina; Dinu, Liviu P.

doi:10.1007/978-3-030-20521-8_57

Cited by 8 publications

(8 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Overfitting can be reduced through training on more than one dataset ( Waseem, Thorne & Bingel, 2018 ; Karan & Šnajder, 2018 ) or transfer learning from a larger dataset ( Uban & Dinu, 2019 ; Alatawi, Alhothali & Moria, 2020 ) and/or a closely related task, such as sentiment analysis ( Uban & Dinu, 2019 ; Cao, Lee & Hoang, 2020 ), yet synthesis in the literature is lacking. More work can be done on comparing different training approaches, and what characteristics of the datasets interact with the effectiveness.…”

Section: Discussionmentioning

confidence: 99%

“…Research on transfer learning from other tasks , such as sentiment analysis, also lacks consistency. Uban & Dinu (2019) pre-trained a classification model on a large sentiment dataset ( https://help.sentiment140.com/ ), and performed transfer learning on the OLID and Kumar datasets. They took pre-training further than the embedding layer, comparing word2vec ( Mikolov et al, 2013 ) to sentiment embeddings and entire-model transfer learning.…”

Section: Obstacles To Generalisable Hate Speech Detectionmentioning

confidence: 99%

See 1 more Smart Citation

Towards generalisable hate speech detection: a review on obstacles and solutions

Yin

Zubiaga

2021

PeerJ Computer Science

View full text Add to dashboard Cite

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Obstacles To Generalisable Hate Speech Detectionmentioning

confidence: 99%

Towards generalisable hate speech detection: a review on obstacles and solutions

Yin

Zubiaga

2021

PeerJ Computer Science

View full text Add to dashboard Cite

show abstract

“…Sahi et al [20] have investigated the automatic detection of hate towards women on Twitter. There exist further studies, which focus on a specific problem, e.g., detecting abusive language [21] or the risks of racial biases in hate speech detection [22]. To automatically detect hate speech, different approaches are used.…”

Section: Challenges and Approaches For Automated Hate Speech Detectionmentioning

confidence: 99%

AI-Assisted and Explainable Hate Speech Detection for Social Media Moderators – A Design Science Approach

Bunde¹

2021

Proceedings of the Annual Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

To date, the detection of hate speech is still primarily carried out by humans, yet there is great potential for combining human expertise with automated approaches. However, identified challenges include low levels of agreement between humans and machines due to the algorithms' missing expertise of, e.g., cultural, and social structures. In this work, a design science approach is used to derive design knowledge and develop an artifact, through which humans are integrated in the process of detecting and evaluating hate speech. For this purpose, explainable artificial intelligence (XAI) is utilized: the artifact will provide explanative information, why the deep learning model predicted whether a text contains hate. Results show that the instantiated design knowledge in form of a dashboard is perceived as valuable and that XAI features increase the perception of the artifact's usefulness, ease of use, trustworthiness as well as the intention to use it.

show abstract

“…Research on transfer learning from other tasks, such as sentiment analysis, also lacks consistency. Uban and Dinu (2019) pre-trained a classification model on a large sentiment dataset 2 , and performed transfer learning on the Zampieri and Kumar datasets. They took pre-training further than the embedding layer, comparing word2vec (Mikolov et al, 2013) to sentiment embeddings and entire-model transfer learning.…”

Section: Existing Solutionsmentioning

confidence: 99%

Towards generalisable hate speech detection: a review on obstacles and solutions

Yin

Zubiaga

2021

Preprint

View full text Add to dashboard Cite

Hate speech is one type of harmful online content which directly attacks or promotes hate towards a group or an individual member based on their actual or perceived aspects of identity, such as ethnicity, religion, and sexual orientation. With online hate speech on the rise, its automatic detection as a natural language processing task is gaining increasing interest. However, it is only recently that it has been shown that existing models generalise poorly to unseen data. This survey paper attempts to summarise how generalisable existing hate speech detection models are, reason why hate speech models struggle to generalise, sums up existing attempts at addressing the main obstacles, and then proposes directions of future research to improve generalisation in hate speech detection. Recent research has raised concerns on the generalisability of existing models (Swamy, Jamatia, & Gambäck, 2019). Despite their impressive performance on their respective test sets, the performance significantly dropped when the models are applied to a different hate speech dataset. This means that the assumption that test data of existing datasets represent the distribution of future cases is not true, and that the generalisation performance of existing models have been severely overestimated (Arango, Pérez, & Poblete, 2020). This lack of generalisability undermines the practical value of these hate speech detection models.

show abstract

On Transfer Learning for Detecting Abusive Language Online

Cited by 8 publications

References 11 publications

Towards generalisable hate speech detection: a review on obstacles and solutions

Towards generalisable hate speech detection: a review on obstacles and solutions

AI-Assisted and Explainable Hate Speech Detection for Social Media Moderators – A Design Science Approach

Towards generalisable hate speech detection: a review on obstacles and solutions

Contact Info

Product

Resources

About