In the ever-evolving landscape of online discourse and political dialogue, the rise of hate speech poses a significant challenge to maintaining a respectful and inclusive digital environment. The context becomes particularly complex when considering the Hindi language—a low-resource language with limited available data. To address this pressing concern, we introduce the CHUNAV dataset—a collection of 11,457 Hindi tweets gathered during assembly elections in various states. CHUNAV is purpose-built for hate speech categorization and the identification of target groups. The dataset is a valuable resource for exploring hate speech within the distinctive socio-political context of Indian elections. The tweets within CHUNAV have been meticulously categorized into “Hate” and “Non-Hate” labels, and further subdivided to pinpoint the specific targets of hate speech, including “Individual”, “Organization”, and “Community” labels (as shown in Figure 1). Furthermore, this paper presents multiple benchmark models for hate speech detection, along with an innovative ensemble and oversampling-based method. The paper also delves into the results of topic modeling, all aimed at effectively addressing hate speech and target identification in the Hindi language. This contribution seeks to advance the field of hate speech analysis and foster a safer and more inclusive online space within the distinctive realm of Indian Assembly Elections.

Social media has become a bedrock for people to voice their opinions worldwide. Due to the greater sense of freedom with the anonymity feature, it is possible to disregard social etiquette online and attack others without facing severe consequences, inevitably propagating hate speech. The current measures to sift the online content and offset the hatred spread do not go far enough. One factor contributing to this is the prevalence of regional languages in social media and the paucity of language flexible hate speech detectors. The proposed work focuses on analyzing hate speech in Hindi-English code-switched language. Our method explores transformation techniques to capture precise text representation. To contain the structure of data and yet use it with existing algorithms, we developed MoH or Map Only Hindi, which means "Love" in Hindi. MoH pipeline consists of language identification, Roman to Devanagari Hindi transliteration using a knowledge base of Roman Hindi words. Finally, it employs the fine-tuned Multilingual Bert and MuRIL language models. We conducted several quantitative experiment studies on three datasets and evaluated performance using Precision, Recall, and F1 metrics. The first experiment studies MoH mapped text's performance with classical machine learning models and shows an average increase of 13% in F1 scores. The second compares the proposed work's scores with those of the baseline models and offers a rise in performance by 6%. Finally, the third reaches the proposed MoH technique with various data simulations using the existing transliteration library. Here, MoH outperforms the rest by 15%. Our results demonstrate a significant improvement in the state-of-the-art scores on all three datasets.

A growing body of work has focused on text classification methods for detecting the increasing amount of hate speech posted online. This progress has been limited to only a select number of highly resourced languages causing detection systems to either under-perform or not exist in limited data contexts. This is mostly caused by a lack of training data, which are expensive to collect and curate in these settings. In this work, we propose a data augmentation approach that addresses the problem of lack of data for online hate speech detection in limited data contexts using synthetic data generation techniques. Given a handful of hate speech examples in a high-resource language such as English, we present three methods to synthesize new examples of hate speech data in a target language that retains the hate sentiment in the original examples but transfers the hate targets. We apply our approach to generate training data for hate speech classification tasks in Hindi and Vietnamese. Our findings show that a model trained on synthetic data performs comparably to, and in some cases outperforms, a model trained only on the samples available in the target domain. This method can be adopted to bootstrap hate speech detection models from scratch in limited data contexts. As the growth of social media within these contexts continues to outstrip response efforts, this work furthers our capacities for detection, understanding, and response to hate speech. Disclaimer: This work contains terms that are offensive and hateful. These, however, cannot be avoided due to the nature of the work.

Sentiment analysis is the most basic NLP task to determine the polarity of text data. There has been a significant amount of work in the area of multilingual text as well. Still hate and offensive speech detection faces a challenge due to inadequate availability of data, especially for Indian languages like Hindi and Marathi. In this work, we consider hate and offensive speech detection in Hindi and Marathi texts. The problem is formulated as a text classification task using the state of the art deep learning approaches. We explore different deep learning architectures like CNN, LSTM, and variations of BERT like multilingual BERT, IndicBERT, and monolingual RoBERTa. The basic models based on CNN and LSTM are augmented with fast text word embeddings. We use the HASOC 2021 Hindi and Marathi hate speech datasets to compare these algorithms. The Marathi dataset consists of binary labels and the Hindi dataset consists of binary as well as more-fine grained labels. We show that the transformer-based models perform the best and even the basic models along with FastText embeddings give a competitive performance. Moreover, with normal hyper-parameter tuning, the basic models perform better than BERT-based models on the fine-grained Hindi dataset.

Hate speech is a specific type of controversial content that is widely legislated as a crime that must be identified and blocked. However, due to the sheer volume and velocity of the Twitter data stream, hate speech detection cannot be performed manually. To address this issue, several studies have been conducted for hate speech detection in European languages, whereas little attention has been paid to low-resource South Asian languages, making the social media vulnerable for millions of users. In particular, to the best of our knowledge, no study has been conducted for hate speech detection in Roman Urdu text, which is widely used in the sub-continent. In this study, we have scrapped more than 90,000 tweets and manually parsed them to identify 5,000 Roman Urdu tweets. Subsequently, we have employed an iterative approach to develop guidelines and used them for generating the Hate Speech Roman Urdu 2020 corpus. The tweets in the this corpus are classified at three levels: Neutral-Hostile, Simple-Complex, and Offensive-Hate speech. As another contribution, we have used five supervised learning techniques, including a deep learning technique, to evaluate and compare their effectiveness for hate speech detection. The results show that Logistic Regression outperformed all other techniques, including deep learning techniques for the two levels of classification, by achieved an F1 score of 0.906 for distinguishing between Neutral-Hostile tweets, and 0.756 for distinguishing between Offensive-Hate speech tweets.

In recent years, monitoring hate speech and offensive language on social media platforms has become paramount due to its widespread usage among all age groups, races, and ethnicities. Consequently, there have been substantial research efforts towards automated detection of such content using Natural Language Processing (NLP). While successfully filtering textual data, no research has focused on detecting hateful content in multimedia data. With increased ease of data storage and the exponential growth of social media platforms, multimedia content proliferates the internet as much as text data. Nevertheless, it escapes the automatic filtering systems. Hate speech and offensiveness can be detected in multimedia primarily via three modalities, i.e., visual, acoustic, and verbal. Our preliminary study concluded that the most essential features in classifying hate speech would be the speaker's emotional state and its influence on the spoken words, therefore limiting our current research to these modalities. This paper proposes the first multimodal deep learning framework to combine the auditory features representing emotion and the semantic features to detect hateful content. Our results demonstrate that incorporating emotional attributes leads to significant improvement over text-based models in detecting hateful multimedia content. This paper also presents a new Hate Speech Detection Video Dataset (HSDVD) collected for the purpose of multimodal learning as no such dataset exists today.

Detecting and classifying instances of hate in social media text has been a problem of interest in Natural Language Processing in the recent years. Our work leverages state of the art Transformer language models to identify hate speech in a multilingual setting. Capturing the intent of a post or a comment on social media involves careful evaluation of the language style, semantic content and additional pointers such as hashtags and emojis. In this paper, we look at the problem of identifying whether a Twitter post is hateful and offensive or not. We further discriminate the detected toxic content into one of the following three classes: (a) Hate Speech (HATE), (b) Offensive (OFFN) and (c) Profane (PRFN). With a pre-trained multilingual Transformer-based text encoder at the base, we are able to successfully identify and classify hate speech from multiple languages. On the provided testing corpora, we achieve Macro F1 scores of 90.29, 81.87 and 75.40 for English, German and Hindi respectively while performing hate speech detection and of 60.70, 53.28 and 49.74 during fine-grained classification. In our experiments, we show the efficacy of Perspective API features for hate speech classification and the effects of exploiting a multilingual training scheme. A feature selection study is provided to illustrate impacts of specific features upon the architecture's classification head.

Abstract Social media has over the years provided a medium for creation and dissemination of opinions and thoughts through online platforms. While it allows users to express their views, sentiments and emotions, some people try to use it to generate and share unpleasant and hateful content. Such content is now referred to as hate speech and it may target an individual, a group, a community, or a country. During the last few years, several techniques have been developed to automatically detect and identify hate speech, offensive and abusive content from social media platforms. However, majority of the studies focused on hate speech detection in English language texts. With social media getting higher penetration across different geographies, there is now a significant amount of content generated in various languages. Though there have been significant advancements in algorithmic approaches for the task, the non-availability of suitable dataset in other languages poses a problem in research advancement in them. Hindi is one such widely spoken language where such datasets are not available. This work attempts to bridge this research gap by presenting a curated and annotated dataset for target-based hate speech (TABHATE) in the Hindi language. The dataset comprises of 2,020 tweets and is annotated by three independent annotators. A multiclass labelling is used where each tweet is labelled as: (i) individual targeting, (ii) community targeting, and (iii) none. Inter annotator agreement is computed. The suitability of dataset is then further explored by applying some standard deep learning and transformer-based models for the task of hate speech detection. The experimental results obtained show that the dataset can be used for experimental work on hate speech detection of Hindi language texts.

Due to the sheer volume of online hate, the AI and NLP communities have started building models to detect such hateful content. Recently, multilingual hate is a major emerging challenge for automated detection where code-mixing or more than one language have been used for conversation in social media. Typically, hate speech detection models are evaluated by measuring their performance on the held-out test data using metrics such as accuracy and F1-score. While these metrics are useful, it becomes difficult to identify using them where the model is failing, and how to resolve it. To enable more targeted diagnostic insights of such multilingual hate speech models, we introduce a set of functionalities for the purpose of evaluation. We have been inspired to design this kind of functionalities based on real-world conversation on social media. Considering Hindi as a base language, we craft test cases for each functionality. We name our evaluation dataset HateCheckHIn. To illustrate the utility of these functionalities , we test state-of-the-art transformer based m-BERT model and the Perspective API.

AbstractThere has been surge in the usage of Internet as well as social media platforms which has led to rise in online hate speech targeted on individual or group. In the recent years, hate speech has resulted in one of the challenging problems that can unfurl at a fast pace on digital platforms leading to various issues such as prejudice, violence and even genocide. Considering the acceptance of Artificial Intelligence (AI) and Natural Language Processing (NLP) techniques in varied application domains, it would be intriguing to consider these techniques for automated hate speech detection. In literature, there have been efforts to recognize and categorize hate speech using varied Machine Learning (ML) and Deep Learning (DL) techniques. Hence, considering the need and provocations for hate speech detection we aim to present a comprehensive review that discusses fundamental taxonomy as well as recent advances in the field of online hate speech identification. There is a significant amount of literature related to the initial phases of hate speech detection. The background section provides a detailed explanation of the previous research. The subsequent section that follows is dedicated to examining the recent literature published from the year 2020 onwards. The paper presents some of the hate speech datasets considered for hate speech detection. Furthermore, the paper discusses different data modalities, namely, textual hate speech detection, multi‐modal hate speech detection and multilingual hate speech detection. Apart from systematic review on hate speech detection, the paper also implement several multi‐label models to compare the performance of hate speech detection by employing classic ML technique namely, Logistic Regression and DL technique namely, Long Short‐Term Memory (LSTM) and a multiclass multi‐label architecture. In the implemented architecture, we have derived two new elements to quantify the hatefulness and intensity of hatred to improve the results for hate speech detection using Indonesian tweet dataset. Empirical Analysis of the model reveals that the implemented approach outperforms and is able to achieve improved results for the underlying dataset.