Search

The prevalence of offensive content on online communication and social media platforms is growing more and more common, which makes its detection difficult, especially in multilingual settings. The term “Offensive Language” encompasses a wide range of expressions, including various forms of hate speech and aggressive content. Therefore, exploring multilingual offensive content, that goes beyond a single language, focus and represents more linguistic diversities and cultural factors. By exploring multilingual offensive content, we can broaden our understanding and effectively combat the widespread global impact of offensive language. This survey examines the existing state of multilingual offensive language detection, including a comprehensive analysis on previous multilingual approaches, and existing datasets, as well as provides resources in the field. We also explore the related community challenges on this task, which include technical, cultural, and linguistic ones, as well as their limitations. Furthermore, in this survey we propose several potential future directions toward more efficient solutions for multilingual offensive language detection, enabling safer digital communication environment worldwide.

The Internet of Things (IoT) stands as one of the most transformative technologies of our era, significantly enhancing the living conditions and operational efficiencies across various domains [...]

THIS ARTICLE USES WORDS OR LANGUAGE THAT IS CONSIDERED PROFANE, VULGAR, OR OFFENSIVE BY SOME READERS. Different types of abusive content such as offensive language, hate speech, aggression, etc. have become prevalent in social media and many efforts have been dedicated to automatically detect this phenomenon in different resource-rich languages such as English. This is mainly due to the comparative lack of annotated data related to offensive language in low-resource languages, especially the ones spoken in Asian countries. To reduce the vulnerability among social media users from these regions, it is crucial to address the problem of offensive language in such low-resource languages. Hence, we present a new corpus of Persian offensive language consisting of 6,000 out of 520,000 randomly sampled micro-blog posts from X (Twitter) to deal with offensive language detection in Persian as a low-resource language in this area. We introduce a method for creating the corpus and annotating it according to the annotation practices of recent efforts for some benchmark datasets in other languages which results in categorizing offensive language and the target of offense as well. We perform extensive experiments with three classifiers in different levels of annotation with a number of classical Machine Learning (ML), Deep learning (DL), and transformer-based neural networks including monolingual and multilingual pre-trained language models. Furthermore, we propose an ensemble model integrating the aforementioned models to boost the performance of our offensive language detection task. Initial results on single models indicate that SVM trained on character or word n-grams are the best performing models accompanying monolingual transformer-based pre-trained language model ParsBERT in identifying offensive vs non-offensive content, targeted vs untargeted offense, and offensive towards individual or group. In addition, the stacking ensemble model outperforms the single models by a substantial margin, obtaining 5% respective macro F1-score improvement for three levels of annotation.

Detecting the underlying human values within arguments is essential across various domains, ranging from social sciences to recent computational approaches. Identifying these values remains a significant challenge due to their vast numbers and implicit usage in discourse. This study explores the potential of emotion analysis as a key feature in improving the detection of human values and information extraction from this field. It aims to gain insights into human behavior by applying intensive analyses of different levels of human values. Additionally, we conduct experiments that integrate extracted emotion features to improve human value detection tasks. This approach holds the potential to provide fresh insights into the complex interactions between emotions and values within discussions, offering a deeper understanding of human behavior and decision making. Uncovering these emotions is crucial for comprehending the characteristics that underlie various values through data-driven analyses. Our experiment results show improvement in the performance of human value detection tasks in many categories.

Social media platforms have surpassed cultural and linguistic boundaries, thus enabling online communication worldwide. However, the expanded use of various languages has intensified the challenge of online detection of hate speech content. Despite the release of multiple Natural Language Processing (NLP) solutions implementing cutting-edge machine learning techniques, the scarcity of data, especially labeled data, remains a considerable obstacle, which further requires the use of semisupervised approaches along with Generative Artificial Intelligence (Generative AI) techniques. This paper introduces an innovative approach, a multilingual semisupervised model combining Generative Adversarial Networks (GANs) and Pretrained Language Models (PLMs), more precisely mBERT and XLM-RoBERTa. Our approach proves its effectiveness in the detection of hate speech and offensive language in Indo-European languages (in English, German, and Hindi) when employing only 20% annotated data from the HASOC2019 dataset, thereby presenting significantly high performances in each of multilingual, zero-shot crosslingual, and monolingual training scenarios. Our study provides a robust mBERT-based semisupervised GAN model (SS-GAN-mBERT) that outperformed the XLM-RoBERTa-based model (SS-GAN-XLM) and reached an average F1 score boost of 9.23% and an accuracy increase of 5.75% over the baseline semisupervised mBERT model.

Recent years brought an exponential growth of social media which revolutionized freedom of speech but significantly increased the propagation of hate speech and hate-based activities. Therefore, constructive countermeasures are necessary to prevent escalating hateful content on online social media. Many recent works target explicit hate speech, but only a few studies have utilized multiple fused features such as sentiment, targets, and emotions as attributes to enhance the detection of hate speech. In general, sentiment features help to discern feelings such as positivity or negativity, and emotion features provide a deeper level of granularity, focusing on a more comprehensive understanding of sensitivities. The aim of this paper is to investigate the significance of incorporating fine-grained emotions as an essential feature in improving the classification of implicit hate speech. First, we analyzed emotion variations of hateful and non-hateful content and explored their major fine-grained emotion discrepancies targeting implicit hateful content. Next, we introduce a multi-task learning approach that integrates emotions and sentiment features to classify implicit expressions of hatred. To evaluate the effectiveness of our multi-task learning approach, we compared it with baseline models using single-task learning approaches. The experimental results show that our multi-task approach outperformed in classifying implicit hate speech compared to the baseline models and demonstrates that fine-grained emotional knowledge decreases the classification error across multiple implicit hate categories.

Automatic detection of abusive online content such as hate speech, offensive language, threats, etc. has become prevalent in social media, with multiple efforts dedicated to detecting this phenomenon in English. However, detecting hatred and abuse in low-resource languages is a non-trivial challenge. The lack of sufficient labeled data in low-resource languages and inconsistent generalization ability of transformer-based multilingual pre-trained language models for typologically diverse languages make these models inefficient in some cases. We propose a meta learning-based approach to study the problem of few-shot hate speech and offensive language detection in low-resource languages that will allow hateful or offensive content to be predicted by only observing a few labeled data items in a specific target language. We investigate the feasibility of applying a meta learning approach in cross-lingual few-shot hate speech detection by leveraging two meta learning models based on optimization-based and metric-based (MAML and Proto-MAML) methods. To the best of our knowledge, this is the first effort of this kind. To evaluate the performance of our approach, we consider hate speech and offensive language detection as two separate tasks and make two diverse collections of different publicly available datasets comprising 15 datasets across 8 languages for hate speech and 6 datasets across 6 languages for offensive language. Our experiments show that meta learning-based models outperform transfer learning-based models in a majority of cases, and that Proto-MAML is the best performing model, as it can quickly generalize and adapt to new languages with only a few labeled data points (generally, 16 samples per class yields an effective performance) to identify hateful or offensive content.

Clickbait can be a spam or an advert which more often provides a link to commercial website and it can also be a headline to news media website which makes money from page views by providing eye-catchy headlines with deceptive news. This paper focuses on the latter definition in order to identify news clickbaits that are published in Twitter. The aim of this work is to use recent Transfer Learning models to detect news clickbaits by adding various configuration changes to the existing models. Based on the author’s knowledge, this is the first attempt to adapt Transfer Learning to classify Clickbaits in social media. In this work we fine-tuned BERT, XLNet and RoBERTa models by integrating novel configuration changes into their default architectures such as model expansion, pruning and data augmentation strategies. Webis Clickbait dataset was used to train these models and the best performed model at the Webit Clickbait competition 2017 was considered as our benchmark. The analyses in this work are mainly focused on eight different scenarios after applying several fine-tuning approaches and model configuration changes to the default Transfer Learning models. The results shown that, our modified Transfer Learning approaches outperformed the considered benchmark. In our experiments, the best performed Transfer Learning model was RoBERTa with the integration of an additional non-linear layer with the hidden output tensor. this configuration has achieved 19.12% more accuracy in compared to the benchmark model for the binary classification. There is no significant performance improvement when each model expanded by adding an extra RNN layer(s). Apart from that, we experimented with another labelled clickbait dataset (Kaggle clickbait challenge) to explore the performance of our fine-tuned models under different scenarios.

The novel coronavirus (COVID-19) pandemic outbreak is drastically shaping and reshaping many aspects of our life, with a huge impact on our social life. In this era of lockdown policies in most of the major cities around the world, we see a huge increase in people and professionals’ engagement in social media. Online Social Networks are playing an important role in news propagation as well as keeping people in contact. At the same time, social media is both a blessing and a curse as the coronavirus infodemic has become a major concern, and is already a topic that needs special attention and further research. In this study, we publish a multilingual coronavirus (COVID-19) Instagram dataset that we have continuously collected during the first wave of the pandemic from 5 January 2020 to 30 May 2020. The dataset contains 25.7K posts, 829K comments, and 3.2M likes in various subjects from different publishers such as ‘public accounts’, ‘fake accounts (bots)’, ‘newsagencies’, ‘influencers’, ‘celebrities’, ‘business pages’, etc. In addition to the dataset, this paper provides an analysis of the behaviour of the publishers. We study the behavioural aspects of the users in terms of their engagement, use of hashtags, activities, reactions as well as a full analysis of the published content related to the COVID-19. We believe this contribution helps the research community to better understand the dynamics behind this phenomenon in Instagram, as one of the major social media.

A framework to consensus opinion model within a networked social group is put forward. The current research in opinion formation within the groups is largely based on the opinion aggregation of each user of the network. However, the consistency of users in aggregation, social power, and the impact of each individual user of the group for opinion formation are not considered. In this paper, we investigate a consensus opinion model in social groups based on the impact of influential users and aggregation methods. In order to reach the consensus model, we aggregate the users' opinions. To maintain consistency, we propagate the opinion through the users to reach an agreement. This propagation will consider the influential users' impacts that have a crucial effect on its process. A novel method is proposed to detect the influential users and opinion propagation based on them to derive the opinion toward the networked social group. In particular, we applied optimism and pessimism scores as the users' personality to discover the influential users in the network. Considering that, we propagate the opinion based on two facts: 1) the impact of influential users, derived by the presence of an extremely confident individual in the network and 2) the impact of neighbors, induced by the presence of the users who have a connection with the current user. Then, we proposed the opinion aggregation of the group induced by the weighted averaging operator and fuzzy techniques. In order to evaluate the validity of the method, we used enormous data sets of Epinions and Etsy which are signed and unsigned, respectively.

Advances in technology have been able to affect all aspects of human life. For example, the use of technology in medicine has made significant contributions to human society. In this article, we focus on technology assistance for one of the most common and deadly diseases to exist, which is brain tumors. Every year, many people die due to brain tumors; based on “braintumor” website estimation in the U.S., about 700,000 people have primary brain tumors, and about 85,000 people are added to this estimation every year. To solve this problem, artificial intelligence has come to the aid of medicine and humans. Magnetic resonance imaging (MRI) is the most common method to diagnose brain tumors. Additionally, MRI is commonly used in medical imaging and image processing to diagnose dissimilarity in different parts of the body. In this study, we conducted a comprehensive review on the existing efforts for applying different types of deep learning methods on the MRI data and determined the existing challenges in the domain followed by potential future directions. One of the branches of deep learning that has been very successful in processing medical images is CNN. Therefore, in this survey, various architectures of CNN were reviewed with a focus on the processing of medical images, especially brain MRI images.

Disparate biases associated with datasets and trained classifiers in hateful and abusive content identification tasks have raised many concerns recently. Although the problem of biased datasets on abusive language detection has been addressed more frequently, biases arising from trained classifiers have not yet been a matter of concern. In this paper, we first introduce a transfer learning approach for hate speech detection based on an existing pre-trained language model called BERT (Bidirectional Encoder Representations from Transformers) and evaluate the proposed model on two publicly available datasets that have been annotated for racism, sexism, hate or offensive content on Twitter. Next, we introduce a bias alleviation mechanism to mitigate the effect of bias in training set during the fine-tuning of our pre-trained BERT-based model for hate speech detection. Toward that end, we use an existing regularization method to reweight input samples, thereby decreasing the effects of high correlated training set' s n-grams with class labels, and then fine-tune our pre-trained BERT-based model with the new re-weighted samples. To evaluate our bias alleviation mechanism, we employed a cross-domain approach in which we use the trained classifiers on the aforementioned datasets to predict the labels of two new datasets from Twitter, AAE-aligned and White-aligned groups, which indicate tweets written in African-American English (AAE) and Standard American English (SAE), respectively. The results show the existence of systematic racial bias in trained classifiers, as they tend to assign tweets written in AAE from AAE-aligned group to negative classes such as racism, sexism, hate, and offensive more often than tweets written in SAE from White-aligned group. However, the racial bias in our classifiers reduces significantly after our bias alleviation mechanism is incorporated. This work could institute the first step towards debiasing hate speech and abusive language detection systems.

The next wave of communication and applications will rely on new services provided by the Internet of Things which is becoming an important aspect in human and machines future. IoT services are a key solution for providing smart environments in homes, buildings, and cities. In the era of massive number of connected things and objects with high growth rate, several challenges have been raised, such as management, aggregation, and storage for big produced data. To address some of these issues, cloud computing emerged to the IoT as Cloud of Things (CoT), which provides virtually unlimited cloud services to enhance the large-scale IoT platforms. There are several factors to be considered in the design and implementation of a CoT platform. One of the most important and challenging problems is the heterogeneity of different objects. This problem can be addressed by deploying a suitable “middleware” which sits between things and applications as a reliable platform for communication among things with different interfaces, operating systems, and architectures. The main aim of this paper is to study the middleware technologies for CoT. Toward this end, we first present the main features and characteristics of middlewares. Next, we study different architecture styles and service domains. Then, we present several middlewares that are suitable for CoT-based platforms and finally, a list of current challenges and issues in the design of CoT-based middlewares is discussed. Keywords: CoT, IoT, Middleware, Fog computing, Cloud