Offensive Detection Research Articles

Suspicionless Stop and Search: You’re Joking, not Another one!

Police stop and search powers have been on the statute book for many years and are widely regarded as being an important tool in the prevention or detection of criminal offences. Whilst most of these powers require that an officer has ‘reasonable grounds to suspect’ that a person is in possession of a prohibited item or article before they may be exercised, the last 30 or so years has seen the emergence of ‘suspicionless’ stop and search powers on the statute book. These have proved to be controversial for a number of reasons, not least because they permit random stop and searches to occur which rarely result in officers finding what they were looking for, and have the potential to strain relations between the police and the communities they serve. Official data also makes clear that they have consistently been used in a racially disproportionate way. Under the Public Order Act 2023, a new suspicionless stop and search power has been created which is exercisable by the police in the context of peaceful protests. The case against it was strongly made in the House of Lords and beyond Parliament. Regrettably, it was not accepted by a government which was also unwilling to amend the legislation so as to enhance the safeguards designed to prevent the misuse of the power.

Wolf Warrior Spreads Superior: The narrative and effectiveness of Chinese public diplomacy behaviours on Twitter

AbstractSince 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.

Syntax-aware Offensive Content Detection in Low-resourced Code-mixed Languages with Continual Pre-training

Social media is a widely used platform that includes a vast amount of user-generated content, allowing the extraction of information about users’ thoughts from texts. Individuals freely express their thoughts on these platforms, often without constraints, even if the content is offensive or contains hate speech. The identification and removal of offensive content from social media are imperative to prevent individuals or groups from becoming targets of harmful language. Despite extensive research on offensive content detection, addressing this challenge in code-mixed languages remains unsolved, characterised by issues such as imbalanced datasets and limited data sources. Most previous studies on detecting offensive content in these languages focus on creating datasets and applying deep neural networks, such as Recurrent Neural Networks (RNNs), or pre-trained language models (PLMs) such as BERT and its variations. Given the low-resource nature and imbalanced dataset issues inherent in these languages, this study delves into the efficacy of the syntax-aware BERT model with continual pre-training for the accurate identification of offensive content and proposes a framework called Cont-Syntax-BERT by combining continual learning with continual pre-training. Comprehensive experimental results demonstrate that the proposed Cont-Syntax-BERT framework outperforms state-of-the-art approaches. Notably, this framework addresses the challenges posed by code-mixed languages, as evidenced by its proficiency on the DravidianCodeMix [10,19] and HASOC 2109 [37] datasets. These results demonstrate the adaptability of the proposed framework in effectively addressing the challenges of code-mixed languages.

Enhancing the fairness of offensive memes detection models by mitigating unintended political bias

Enhancing the fairness of offensive memes detection models by mitigating unintended political bias

The Effectiveness of Electronic Declaration Monitoring in the Detection and Investigation of Corruption Offences in the Judicial System

The Effectiveness of Electronic Declaration Monitoring in the Detection and Investigation of Corruption Offences in the Judicial System

Unsupervised offensive speech detection for multimedia based on multilingual BERT

Unsupervised offensive speech detection for multimedia based on multilingual BERT

A comparison of text preprocessing techniques for hate and offensive speech detection in Twitter

A comparison of text preprocessing techniques for hate and offensive speech detection in Twitter

A feature fusion and detection approach using deep learning for sentimental analysis and offensive text detection from code-mix Malayalam language

Social media has efficiently turned out to be a major source of digital marketing and communication. As social media platforms set up free manifestation of facts and thoughts in images, video, and text, there is a widespread requirement to screen them to secure groups and individuals from being exposed to offensive content that specifically targets them. In social media communications, limitless practice of vocabulary complicates the code-mixed text processing. The most important aspiration of this research is to analyze the sentiments and classify the code-mixed social media posts/comments in the Dravidian language: Malayalam using deep learning. In this study, initial phase is the ALBERT tokenization that splits the words in the comments. Subsequently, to extract significant features from the dataset, feature extraction is carried out. Following feature extraction, feature fusion takes place using the Shannon entropy and Recurrent Neural Network (RNN) model. Finally, sentimental analysis and offensive text identification are conducted using the Hierarchical Attention Network (HAN). The experimentation of the proposed Feature fusion + HAN technique is performed by the metrics, namely accuracy, sensitivity, and specificity, where the accuracy attained by the proposed model is 0.956, the sensitivity is 0.939, and the specificity is 0.978.

PREVENTION OF LEGALISATION (LAUNDERING) OF CRIMINAL PROCEEDS IN THE MECHANISM OF ENSURING ECONOMIC SECURITY OF THE STATE

The purpose of the article is to consider the nature and content of economic security as a component of national security of the state; to determine the peculiarities of implementation of a criminal illegal intention to commit legalisation (laundering) of criminally obtained property; to establish the main destructive features of legalisation (laundering) of criminally obtained property and their correlation with economic security of the state; to define the actual means of prevention of legalisation (laundering) of criminally obtained property in the mechanism of ensuring economic security of the state. A scientific discussion of the problem of ensuring the economic security of the country by reducing the risks of legalisation (laundering) of criminally acquired property is presented. The regulatory and programme documents in the field of ensuring economic security have been characterised. The nature and content of economic security as a component of the state's national security were determined. Statistical reporting of law enforcement bodies in terms of detection, investigation and prevention of predicate offences and direct legalisation (laundering) of criminally acquired property was analysed. The authors emphasise the need to improve the system of timely detection of economic risks and threats in order to avoid regulatory, supervisory, organisational and managerial gaps that facilitate both the commission of predicate criminal offences and direct legalisation of the proceeds of crime. Results. It has been established that the essence and content of economic security as a component of the state's national security consists in such functional features as: 1) ensuring the implementation of full cycle production as a means of increasing the level of resource efficiency of the economy; 2) providing foreign and domestic trade infrastructure as a means of increasing the institutional economic capacity of the state; 3) ensuring the investment attractiveness of the state as a means of improving the foreign and domestic economic policy of the state (investment and reinvestment processes); 4) ensuring economic stimulation for the systematisation of innovative renewal of the products of the activities of business entities; 5) ensuring control over labour migration in order to retain qualified specialists; 6) ensuring protection against economic crimes; 7) ensuring the development and competitiveness of the stock market. The specifics of the implementation of the criminal illegal intention to commit legalisation (laundering) of criminally obtained property have been differentiated: 1) multi-stage, which implies the presence of a complex mechanism of criminally illegal behaviour, often aimed at achieving several goals united by a single intention; 2) a list of possible predicate offences, which are fundamental in terms of property legalisation (laundering); 3) committing a criminal offence as part of an organised group/criminal organisation; 4) involving partners in crime with interregional and cross-border links; 5) ensuring the latency of predicate offences through legalisation (laundering) of criminally acquired property. Modern methods of prevention of legalisation (laundering) of criminally acquired property in the mechanism of ensuring economic security of the state have been proposed: 1) differentiation of legalisation (laundering) of criminally acquired property as an economic and criminal offence based on the object of intervention in/outside the predicate socially dangerous acts; 2) improvement of financial monitoring and registration procedures related to the acquisition of property rights and other economic transactions; 3) a fundamental departure from the generalisation of measures aimed at de-shadowing the economy (taking into account the object of de-shadowing); 4) elimination of shortcomings in economic procedural legislation, which often creates a field for fictitious bankruptcy; 5) elimination of shortcomings in criminal legislation, by revising the sanction of Art. 209 of the Criminal Code of Ukraine.

Arabic cyberbullying detecting using ensemble deep learning technique

<span lang="EN-US">There has been a huge growth in recent years interest in studies on abusive language and cyberbullying detection due to its effects on both individual victims and societies. Hate speech, bullying, racism, aggressive content, harassment and other forms of abuse have all significantly increased as a result of Facebook, Instagram, and other social media platforms (SMPs). Since there is a significant need to detect, control, and prohibit the circulation of offensive content on social networking sites, we undertook this study to automate the identification of abusive language or cyberbullying. Arabic data set is balanced and will be used in the offensive detection process. Recently, ensemble machine learning has been used to increase the effectiveness of categorization models. Arabic detection is more precise given that each spatial feature text can make references to every other contextual piece of information. The authors utilized a model that merged convolutional neural network (CNN) with bidirectional long short-term memory (Bi-LSTM) and inverse document frequency gated recurrent unit (GRU) in a hybrid fashion without any post-processing. Our work outperformed every other publicly released cutting-edge ensemble model in the specifications of the official deep learning challenge. The findings indicate that the three-layer inverse document frequency long short-term memory (LSTM) classifier surpassed other classifiers in accuracy with a score of 92.75% compared to different algorithms.</span>

‘You don’t know what you’ve got ‘til it's gone’: Police retention of investigative materials

The seizure, safe handling and secure storage of evidence during criminal investigations is pivotal to the successful detection and prosecution of offences. The safe retention of materials after an investigation closes, even post-conviction, is also critical both to the ability of the appellate system to function and for ‘cold cases’ to be reviewed. Yet despite periodic high-profile instances in which failures to ensure the integrity and retention of evidence have occurred, there is a reticence to admit to evidence being lost; in particular, to consider the issue on an aggregate rather than a case-by-case basis. This lack of transparency is compounded by an accountability deficit, in which the issue falls through regulatory gaps. Provoked to locate proof of lost evidence, this article examines police retention of investigative materials, and details the results of surveys undertaken with lawyers and miscarriage of justice campaigners ( n = 65) and police ( n = 87), and 21 interviews with serving and retired police officers. We identify ‘weak links’ in the evidence chain that need strengthening, including the enhancement of training, improved physical packaging and storage facilities, sufficient training of police staff and the creation of specialised ‘exhibits’ roles. Further to practical resource demands, the issue requires accessible data so that the problem can be acknowledged, and accountability and oversight mechanisms created to assure processes.

Detection of offensive terms in resource-poor language using machine learning algorithms.

The use of offensive terms in user-generated content on different social media platforms is one of the major concerns for these platforms. The offensive terms have a negative impact on individuals, which may lead towards the degradation of societal and civilized manners. The immense amount of content generated at a higher speed makes it humanly impossible to categorise and detect offensive terms. Besides, it is an open challenge for natural language processing (NLP) to detect such terminologies automatically. Substantial efforts are made for high-resource languages such as English. However, it becomes more challenging when dealing with resource-poor languages such as Urdu. Because of the lack of standard datasets and pre-processing tools for automatic offensive terms detection. This paper introduces a combinatorial pre-processing approach in developing a classification model for cross-platform (Twitter and YouTube) use. The approach uses datasets from two different platforms (Twitter and YouTube) the training and testing the model, which is trained to apply decision tree, random forest and naive Bayes algorithms. The proposed combinatorial pre-processing approach is applied to check how machine learning models behave with different combinations of standard pre-processing techniques for low-resource language in the cross-platform setting. The experimental results represent the effectiveness of the machine learning model over different subsets of traditional pre-processing approaches in building a classification model for automatic offensive terms detection for a low resource language, i.e., Urdu, in the cross-platform scenario. In the experiments, when dataset D1 is used for training and D2 is applied for testing, the pre-processing approach named Stopword removal produced better results with an accuracy of 83.27%. Whilst, in this case, when dataset D2 is used for training and D1 is applied for testing, stopword removal and punctuation removal were observed as a better preprocessing approach with an accuracy of 74.54%. The combinatorial approach proposed in this paper outperformed the benchmark for the considered datasets using classical as well as ensemble machine learning with an accuracy of 82.9% and 97.2% for dataset D1 and D2, respectively.

Comparative performance of ensemble machine learning for Arabic cyberbullying and offensive language detection

Since cyberbullying impacts both individual victims and entire society, research on abusive language and its detection has attracted attention in recent years. Because social media sites like Facebook, Instagram, Twitter, and others are so widely accessible, hate speech, bullying, sexism, racism, aggressive material, harassment, poisonous comments, and other types of abuse have all substantially increased. Due to the critical requirement to detect, regulate, and limit the spread of harmful content on social networking sites, we conducted this study to automate the detection of offensive language or cyberbullying. We created a new Arabic balanced data set to be used in the offensive detection process because having a balanced data set for a model would result in improved accuracy models. Recently, the performance of single classifiers has been improved using ensemble machine learning. The purpose of this study is to examine the effectiveness of several single and ensemble machine learning algorithms in identifying Arabic text that contains foul language and cyberbullying. Applying them to three Arabic datasets, we have selected three machine learning classifiers and three ensemble models for this aim. Two of them are offensive datasets that are readily accessible in the public, while the third one was created. The results showed that the single learner machine learning strategy is inferior to the ensemble machine learning methodology. Voting performs is the best performing trained ensemble machine learning classifier, outperforming the best single learner classifier (65.1%, 76.2%, and 98%) for the same datasets with accuracy scores of (71.1%, 76.7%, and 98.5%) for each of the three datasets used. Finally, we improve the voting technique’s performance through hyperparameter tuning on the Arabic cyberbullying data set.

Problems of interaction in the investigation of corruption crimes in the field of official activity and professional activity related to the provision of public services

Modern realities and the state of the criminogenic situation in the country require fundamentally new approaches to the interaction of law enforcement agencies. Today, the interaction of the investigator with other bodies and units is more and more actively carried out at all possible levels, departmental, interdepartmental, international. Interaction in the detection and investigation of criminal offences is a rather complex and comprehensive form that involves a number of interrelated procedural and operational search actions and measures aimed at maximising the achievement of the goals of the criminal process. Cooperation in the investigation of crimes, in particular in the field of official activity, is one of the methods of combating crime, which consists in the coordination and combination in order and on the basis of the law, the efforts of several persons, divisions, bodies, etc. in order to ensure the optimization of the pre-trial investigation of criminal offenses. The specified activity is objectively necessary and often begins from the moment of checking the received information about the commission of a crime. This activity by its nature is very complex, which finds its expression in its phasing. Among the latter, we suggest making a decision about the necessity and expediency of interaction, choosing the form and methods of interaction, procedural consolidation of the decision, direct interaction, and registration of the results of the indicated cooperation. However, the structural clarity of the interaction does not relieve it of the difficulties faced by its participants. This situation is mainly due to the lack of clear regulation of issues related to the organisation and implementation of cooperation in the Criminal Procedure Code of Ukraine. In connection with this, it would be expedient to supplement the specified legal act with a separate section, which would define the concepts, principles, types, forms, methods and procedure of interaction during criminal proceedings.
 Having analysed the legal acts, the author concludes that there is no clear regulation of such interaction.

Ar-PuFi: A short-text dataset to identify the offensive messages towards public figures in the Arabian community

The fight against offensive speech on the Internet necessitates increased efforts from linguistic analysis and artificial intelligence perspectives to develop countermeasures and preventive methods. Reliable predictions can only be obtained if these methods are exposed to a representative sample of the domain or environment under consideration. Datasets serve as the foundation for significant developments in this field because they are the main means of obtaining appropriate instances that reveal the multiple and varied faces of the offensive speech phenomenon. In this sense, we present Ar-PuFi, a dataset of offensive speech towards Public Figures in the Arabian community. With 24,071 comments collected from TV interviews with Egyptian celebrities belonging to six domains of public interest, Ar-PuFi is currently the largest Arabic dataset in terms of its category and size. The examples were annotated by three native speakers over the course of two months and are provided with two-class and six-class annotations based on the presence or absence of explicit and implicit offensive content. We evaluated the performance of a diverse set of classification models employing several text representations of actual examples (e.g., N-gram, TF/IDF, AraVec, and fastText), and AraBERT achieved the baseline for the new dataset in both offensive detection and group classification. Additionally, we apply the Pointwise Mutual Information (PMI) technique to comments within the target domain in order to derive a lexicon of offensive terms associated with each domain of ArPuFi. We further explored whether active learning (AL) or meta-learning (ML) frameworks could be used to reduce the labeling effort required for our dataset without affecting prediction quality and found that, though AL can reduce the amount of data annotations by 10% over the ML approach, neither approach requires less than about 70% of the full dataset to achieve baseline performance. Finally, we took advantage of the availability of relevant datasets and conducted a cross-domain experiment to back up our claims not only about the uniqueness of our dataset but also about the difficulty of adapting Arabic dialects against one another.

Improved GBS-YOLOv5 algorithm based on YOLOv5 applied to UAV intelligent traffic

As the road traffic situation becomes complex, the task of traffic management takes on an increasingly heavy load. The air-to-ground traffic administration network of drones has become an important tool to promote the high quality of traffic police work in many places. Drones can be used instead of a large number of human beings to perform daily tasks, as: traffic offense detection, daily crowd detection, etc. Drones are aerial operations and shoot small targets. So the detection accuracy of drones is less. To address the problem of low accuracy of Unmanned Aerial Vehicles (UAVs) in detecting small targets, we designed a more suitable algorithm for UAV detection and called GBS-YOLOv5. It was an improvement on the original YOLOv5 model. Firstly, in the default model, there was a problem of serious loss of small target information and insufficient utilization of shallow feature information as the depth of the feature extraction network deepened. We designed the efficient spatio-temporal interaction module to replace the residual network structure in the original network. The role of this module was to increase the network depth for feature extraction. Then, we added the spatial pyramid convolution module on top of YOLOv5. Its function was to mine small target information and act as a detection head for small size targets. Finally, to better preserve the detailed information of small targets in the shallow features, we proposed the shallow bottleneck. And the introduction of recursive gated convolution in the feature fusion section enabled better interaction of higher-order spatial semantic information. The GBS-YOLOv5 algorithm conducted experiments showing that the value of mAP@0.5 was 35.3% and the mAP@0.5:0.95 was 20.0%. Compared to the default YOLOv5 algorithm was boosted by 4.0% and 3.5%, respectively.

Attention Mechanism Architecture for Arabic Sentiment Analysis

This article tackles the problem of sentiment analysis in the Arabic language where a new deep learning model has been put forward. The proposed model uses a hybrid bidirectional gated recurrent unit (BiGRU) and bidirectional long short-term memory (BiLSTM) additive-attention model where the Bidirectional GRU/LSTM reads the individual sentence input from left to right and vice versa, enabling the capture of the contextual information. However, the model is trained on two types of embeddings: FastText and local learnable embeddings. The BiLSTM and BiGRU architectures are put into competition to identify the best hyperparameter set for the model. The developed model has been tested on three large-scale commonly employed Arabic sentiment dataset: large-scale Arabic Book Reviews Dataset (ABRD), Hotel Arabic-Reviews Dataset (HARD), and Books Reviews in the Arabic Dataset (BRAD). The testing results demonstrate that our model outperforms both the baseline models and the state-of-the-art models reported in the original references of these datasets, achieving accuracy scores of 98.6%, 96.19%, 95.65% for LARB, HARD, and BRAD, respectively. Furthermore, to demonstrate the generalization capabilities of our model, the performances of the model have been evaluated on three other natural language processing tasks: news categorization, offensive speech detection, and Russian sentiment analysis. The results demonstrated the developed model is language- and task-independent, which offers new perspectives for the application of the developed models in several other natural language processing challenges.

Hope speech detection in Spanish

In recent years, systems have been developed to monitor online content and remove abusive, offensive or hateful content. Comments in online social media have been analyzed to find and stop the spread of negativity using methods such as hate speech detection, identification of offensive language or detection of abusive language. We define hope speech as the type of speech that is able to relax a hostile environment and that helps, gives suggestions and inspires for good to a number of people when they are in times of illness, stress, loneliness or depression. Detecting it automatically, in order to give greater diffusion to positive comments, can have a very significant effect when it comes to fighting against sexual or racial discrimination or when we intend to foster less bellicose environments. In this article we perform a complete study on hope speech, analyzing existing solutions and available resources. In addition, we have generated a quality resource, SpanishHopeEDI, a new Spanish Twitter dataset on LGBT community, and we have conducted some experiments that can serve as a baseline for further research.

Information about computer handwriting: The challenges of finding a balance between individual liberty and state security

The present study is devoted to the introduction of biometric identification systems in law enforcement. The author pays special attention to the phenomenon of computer handwriting, examining in detail the constitutional and legal aspects of the use of knowledge about it. For this purpose a wide methodological apparatus is used: comparative-legal method, method of analogy, deduction, method of modelling and others. The Russian practice and potential for further development are compared with the experience of foreign countries both in the regulation of general issues of biometric personal data and in situations of legal regulation of computer handwriting data. The author investigates how to simultaneously solve the problems of detection and investigation of offences in which printed texts act as evidentiary materials, and at the same time preserve human rights to privacy, personal and family secrecy, anonymity on the Internet. It is noted that computer handwriting is a rather young phenomenon for legal science; there is still no in-depth study of it from the perspective of this field of knowledge. This paper is only a beginning in understanding the topic under consideration. It raises basic questions: whether it is permissible to use knowledge about computer handwriting, how it should be protected and where it should be stored, whether it is necessary to limit the existing commercial interest in its use. As a result, the author comes to the following conclusions: it is permissible and necessary to collect information on computer handwriting, as this will reduce infringements on the natural rights of man and citizen. For such collection it is necessary to use a centralised state system in which all the requirements for the security of critical information infrastructure are implemented. However, before such a system is put into operation, it is necessary to support commercial companies in relevant developments and oblige them to provide information about users' computer handwriting at the request of authorised persons.

PCROD: Context Aware Role based Offensive Detection using NLP/ DL Approaches

With the increased use of social media many people misuse online platforms by uploading offensive content and sharing the same with vast audience. Here comes controlling of such offensive contents. In this work we concentrate on the issue of finding offensive text in social media. Existing offensive text detection systems treat weak pejoratives like ‘idiot‘ and extremely indecent pejoratives like ‘f***‘ as same as offensive irrespective of formal and informal contexts . In fact the weakly pejoratives in informal discussions among friends are casual and common which are not offensive but the same can be offensive when expressed in formal discussions. Crucial challenges to accomplish the task of role based offensive detection in text are i) considering the roles while classifying the text as offensive or not i) creating a contextual datasets including both formal and informal roles. To tackle the above mentioned challenges we develop deep neural network based model known as context aware role based offensive detection(CROD). We examine CROD on the manually created dataset that is collected from social networking sites. Results show that CROD gives better performance with RoBERTa with an accuracy of 94% while considering the context and role in data specifics.