Offensive communications have invaded social media content. One of the most effective solutions to cope with this problem is using computational techniques to discriminate offensive content. ...Moreover, social media users are from linguistically different communities. This study aims to tackle the Multilingual Offensive Language Detection (MOLD) task using transfer learning models and the fine-tuning phase. We propose an effective approach based on the Bidirectional Encoder Representations from Transformers (BERT) that has shown great potential in capturing the semantics and contextual information within texts. The proposed system consists of several stages: (1) Preprocessing, (2) Text representation using BERT models, and (3) Classification into two categories: Offensive and non-offensive. To handle multilingualism, we explore different techniques such as the joint-multilingual and translation-based ones. The first consists in developing one classification system for different languages, and the second involves the translation phase to transform all texts into one universal language then classify them. We conduct several experiments on a bilingual dataset extracted from the Semi-supervised Offensive Language Identification Dataset (SOLID). The experimental findings show that the translation-based method in conjunction with Arabic BERT (AraBERT) achieves over 93% and 91% in terms of F1-score and accuracy, respectively.
Abstract Since 2013, when President Xi Jinping pioneered the concept of "telling China's stories well," the number of senior Chinese diplomats and state-affiliated media accounts on Twitter has ...increased. Different from the vague and evasive diplomatic parlance, some diplomats defend the policies of China in a relatively aggressive way, even sometimes resulting in online disputes with foreign politicians. They are labeled as "wolf-warrior diplomats," a term coined from the record-breaking Chinese nationalist action movie series Wolf Warrior. This paper investigates the effectiveness of China's "wolf warrior diplomacy" on audience engagement on Twitter and significant factors impacting communication effectiveness. Through the utilization of advanced offensive and humor detection algorithms, counterintuitively, this study finds that the wolf-warrior tweets improve Twitter audience engagement, though prior research pointed out that these tweets may bring out adverse feelings in some audiences. Moreover, it also unveils that providing more information and posting humorously on Chinese diplomatic Tweets can enhance their reach and dissemination.
•A thorough review of the techniques, algorithms, datasets, and tasks for offensive language detection in Dravidian languages.•Novel MPNet and CNN fusion technique for offensive language detection in ...low-resource Dravidian languages.•An extensive evaluation of benchmark datasets with positive results.
Social media has effectively replaced traditional forms of communication and marketing. As these platforms allow for the free expression of ideas and facts through text, images, and videos, there exists a significant need to screen them to safeguard people and organisations from objectionable information directed at them. Our work aims to categorise code-mixed social media comments and posts in Tamil, Malayalam, and Kannada into offensive or not offensive at different levels. We present a multilingual MPNet and CNN fusion model for detecting offensive language content directed at an individual (or group) in low-resource Dravidian languages at different levels. Our model is capable of handling data that has been code-mixed, such as Tamil and Latin scripts. The model was successfully validated on the datasets, achieving offensive language detection results better than those of other baseline models with weighted average F1-score of 0.85, 0.98, and 0.76, and performed better than the baseline models EWDT, and EWODT by 0.02, 0.02, 0.04 for Tamil, Malayalam, and Kannada respectively.
The pervasiveness of offensive content in social media has become an important reason for concern for online platforms. With the aim of improving online safety, a large number of studies applying ...computational models to identify such content have been published in the last few years, with promising results. The majority of these studies, however, deal with high-resource languages such as English due to the availability of datasets in these languages. Recent work has addressed offensive language identification from a low-resource perspective, exploring data augmentation strategies and trying to take advantage of existing multilingual pretrained models to cope with data scarcity in low-resource scenarios. In this work, we revisit the problem of low-resource offensive language identification by evaluating the performance of multilingual transformers in offensive language identification for languages spoken in India. We investigate languages from different families such as Indo-Aryan (e.g., Bengali, Hindi, and Urdu) and Dravidian (e.g., Tamil, Malayalam, and Kannada), creating important new technology for these languages. The results show that multilingual offensive language identification models perform better than monolingual models and that cross-lingual transformers show strong zero-shot and few-shot performance across languages.
•We develop a measure of concern for political correctness (CPC) with two factors.•The emotion factor (PC-E) is related to lower emotional well-being.•The activism factor (PC-A) is related to ...frequent arguments and losing friends.•The activism factor (PC-A) is related to a more liberal and activist identity.•Criterion validation shows PC-E predicts response to politically incorrect humor.
The transformation of common language toward inclusion of all people is the mechanism by which many aim to alter attitudes and beliefs that stand in the way of more meaningful social change. The term for this motivated concern for language is “political correctness” or “PC.” The current project seeks to introduce a new tool for investigations into this phenomenon, the concern for political correctness (CPC) scale. CPC assesses individual differences in concern for politically correct speech. Exploratory and confirmatory structural equation modeling showed consistent factor structure of the two subscales; an emotion subscale measuring negative emotional response to hearing politically incorrect language, and an activism subscale measuring a willingness to correct others who use politically incorrect language. Correlational analyses suggested that concern for political correctness is associated with more liberal beliefs and ideologies and less right-wing authoritarianism. The emotion subscale was also found to be associated with lower emotional well-being and the activism subscale with more frequent arguments. Laboratory-based criterion validation studies indicated that the two subscales predicted negative reactions to politically incorrect humor.
The present paper focuses on the presentation and discussion of aspects of OFFENSIVE LANGUAGE linguistic annotation, including the creation, annotation practice, curation, and evaluation of an ...OFFENSIVE LANGUAGE annotation taxonomy scheme, that was first proposed in Lewandowska-Tomaszczyk et al. (2021). An extended offensive language ontology comprising 17 categories, structured in terms of 4 hierarchical levels, has been shown to represent the encoding of the defined offensive language schema, trained in terms of non-contextual word embeddings – i.e., Word2Vec and Fast Text, and eventually juxtaposed to the data acquired by using a pair wise training and testing analysis for existing categories in the HateBERT model (Lewandowska-Tomaszczyk et al. submitted). The study reports on the annotation practice in WG 4.1.1. Incivility in media and social media in the context of COST Action CA 18209 European network for Web-centred linguistic data science (Nexus Linguarum) with the INCEpTION tool (https://github.com/inception-project/inception) – a semantic annotation platform offering assistance in the annotation. The results partly support the proposed ontology of explicit offense and positive implicitness types to provide more variance among widely recognized types of figurative language (e.g., metaphorical, metonymic, ironic, etc.). The use of the annotation system and the representation of linguistic data were also evaluated in a series of the annotators’ comments, by means of a questionnaire and an open discussion. The annotation results and the questionnaire showed that for some of the categories there was low or medium inter-annotator agreement, and it was more challenging for annotators to distinguish between category items than between aspect items, with the category items offensive, insulting and abusive being the most difficult in this respect. The need for taxonomic simplification measures on the basis of these results has been recognized for further annotation practices.
As social media platforms offer a medium for opinion expression, social phenomena such as hatred, offensive language, racism, and all forms of verbal violence have increased spectacularly. These ...behaviors do not affect specific countries, groups, or communities only, extending beyond these areas into people’s everyday lives. This study investigates offensive and hate speech on Arab social media to build an accurate offensive and hate speech detection system. More precisely, we develop a classification system for determining offensive and hate speech using a multi-task learning (MTL) model built on top of a pre-trained Arabic language model. We train the MTL model on the same task using cross-corpora representing a variation in the offensive and hate context to learn global and dataset-specific contextual representations. The developed MTL model showed a significant performance and outperformed existing models in the literature on three out of four datasets for Arabic offensive and hate speech detection tasks.
The proliferation of harmful content on social media affects a large part of the user community. Therefore, several approaches have emerged to control this phenomenon automatically. However, this is ...still a quite challenging task. In this paper, we explore the offensive language as a particular case of harmful content and focus our study in the analysis of keywords in available datasets composed of offensive tweets. Thus, we aim to identify relevant words in those datasets and analyze how they can affect model learning. For keyword extraction, we propose an unsupervised hybrid approach which combines the multi-head self-attention of BERT and a reasoning on a word graph. The attention mechanism allows to capture relationships among words in a context, while a language model is learned. Then, the relationships are used to generate a graph from what we identify the most relevant words by using the eigenvector centrality. Experiments were performed by means of two mechanisms. On the one hand, we used an information retrieval system to evaluate the impact of the keywords in recovering offensive tweets from a dataset. On the other hand, we evaluated a keyword-based model for offensive language detection. Results highlight some points to consider when training models with available datasets.
PDF
