The proliferation of hate speech on digital platforms has evolved from primarily text-based expressions to more sophisticated multimodal forms, combining offensive language with harmful images. With the advent of AI-generated content, this phenomenon has become more complex, making it increasingly difficult for conventional hate speech detection models to address it adequately. Adding to this challenge is the emergence of adversarial attacks, which deliberately manipulate inputs to mislead detection models, further exposing their vulnerabilities and reducing their effectiveness. This paper investigates the vulnerabilities of multimodal classifiers in detecting AI-generated multimodal hate speech when exposed to adversarial attacks and focuses on crafting adversarial examples that simultaneously exploit textual and visual components. Applying perturbations in the direction of the model’s loss function gradient demonstrates how minimal, imperceptible changes to both text and image inputs can deceive state-of-the-art classifiers. A novel technique, ADVattack, was introduced for generating adversarial examples using cross-model feature manipulation in this multimodal context and analyzing current models’ susceptibility to such attacks. The results indicate a significant gap in the robustness of existing multimodal hate speech detectors, highlighting the need for improved defense mechanisms against these sophisticated adversarial threats.

The issue of sexism in online content has recently been a significant concern. With the increasing number of online interactions and the rise of social media platforms, the need for automated techniques to identify and classify sexism has become more critical than ever. This paper addresses this problem by fine-tuning deep-learning models for sexism classification using “MultiHate”. It is a comprehensive dataset created by curating ten different datasets on sexism. The dataset consists of 1.76 M English texts labelled as sexist and not sexist, then fine-tuned two deep learning models, Convolutional Neural Networks-Bidirectional Long Short-Term Memory and Generative Pre-trained Transformer 2, which accurately detect and classify sexism. A comparative analysis has been conducted on several machine learning and deep learning models using the MultiHate dataset. Investigation reveals that the Generative Pre-trained Transformer 2 model outperforms other models with an accuracy of 92%, while the Convolutional Neural Networks-Bidirectional Long Short-Term Memory model achieved an accuracy of 90% using precision, recall, and F1 scores as performance metrics. The models’ performances are promising, indicating that automated techniques can be employed to classify sexist content effectively. A comprehensive error analysis of the models’ performance has been presented, highlighting their limitations and challenges. The computational time required for training and testing the models is a significant challenge, especially for larger datasets.

The rise of multimodal hate speech, which combines text and visual elements, poses significant challenges for online content moderation. Traditional detection models often focus on single modalities and struggle with AI-generated content that is contextually nuanced and semantically complex. These limitations lead to suboptimal performance, as existing frameworks are not robust enough to handle the evolving nature of hate speech across diverse contexts and datasets. An integrated approach that captures the interplay between text and images is needed for more accurate identification. This paper introduces a novel MultiLate classifier designed to synergistically integrate text and image modalities for robust hate speech detection to address these challenges. The textual component employs a CNN-BiLSTM architecture, augmented by a feature fusion pipeline incorporating Three W’s Question Answering and sentiment analysis. For the image modality, the classifier utilizes a pre-trained ResNet50 architecture alongside Diffusion Attention Attribution Maps to generate pixel-level heatmaps, highlighting salient regions corresponding to contextually significant words. These heatmaps are selectively processed to enhance both classification accuracy and computational efficiency. The extracted features from both modalities are then fused to perform comprehensive multimodal classification. Extensive evaluations of the MULTILATE and MultiOFF datasets demonstrate the efficacy of the proposed approach. Comparative analysis against state-of-the-art models underscores the robustness and generalization capability of the MultiLate classifier. The proposed framework enhances detection accuracy and optimizes computational resource utilization, signifi-cantly advancing multimodal hate speech classification.

This study proposes the results of the low-resource Indic language translation task organized in collaboration with the Tenth Conference on Machine Translation (WMT) 2025. In this workshop, participants were required to build and develop machine translation models for the seven language pairs, which were categorized into two categories. Category 1 is moderate training data available in languages i.e English-Assamese, English-Mizo, English-Khasi, English-Manipuri and English-Nyishi. Category 2 has very limited training data available in languages, i.e English-Bodo and English-Kokborok. This task leverages the enriched IndicNE-corp1.0 dataset, which consists of an extensive collection of parallel and monilingual corpora for north eastern Indic languages. The participant results were evaluated using automatic machine translation metrics, including BLEU, TER, ROUGE-L, ChrF, and METEOR. Along with those metrics, this year's work also includes Cosine similarity for evaluation, which captures the semantic representation of the sentence to measure the performance and accuracy of the models. This work aims to promote innovation and advancements in low-resource Indic languages.

Sexism in social networks manifests in various forms, from blatant misogyny to subtle, implicit biases, presenting a significant societal challenge that necessitates effective detection and mitigation strategies. Addressing this issue involves participation in the EXIST 2024 tasks, a competition designed to advance the identification of sexist content in social media. This year's contest includes both traditional text-based data from tweets and an innovative meme dataset, incorporating both images and text. The approach leverages sophisticated models to analyze these multimodal inputs. For textual modalities, a Convolutional Neural Network - Bidirectional Long Short-Term Memory model is employed to discern sexist language and tweet behaviours. For image modalities, a combination of Residual Network 50 and text-based analysis is utilized to detect and interpret sexist elements within memes. Both models undergo hyperparameter tuning and k-fold cross-validation to ensure robustness and accuracy. Preliminary results indicate that integrating these methods enhances the precision and effectiveness of sexism detection, providing a comprehensive tool for identifying and addressing sexist content in diverse social media formats.

In the contemporary landscape, two-wheelers have emerged as the predominant mode of transportation, despite their inherent risk due to limited protection. Disturbing data from 2020 reveals a daily toll of 304 lives lost in India in road accidents involving two-wheeler riders without helmets, emphasizing the urgent need for safety measures. Recognizing the crucial role of helmets in mitigating risks, governments have made riding without one a punishable offense, employing manual strategies for enforcement with limitations in speed and weather conditions. In today’s world of advancing technology, we can leverage the power of computer vision and deep learning to tackle this problem. This can eliminate the need for constant human surveillance to be kept on riders and can automate this process, thus enforcing law and order as well as making this process efficient. Our proposed solution utilizes video surveillance and the YOLOv8 deep learning model for automatic helmet detection. The system employs pure machine learning to identify helmet types with minimal computation cost by utilizing various image processing algorithms. Once the helmet-less person is detected, the number plate corresponding to the rider’s motorcycle is also detected and extracted using computer vision techniques. This number plate is then stored in a database thus allowing further intervention to be done in this matter by the authorities to ensure penalties and enforce safety rules properly. The model developed achieves an overall accuracy score of 93.6% on the testing data, thus showcasing good results on diverse datasets.

This paper presents the results of the low-resource Indic language translation task, organized in conjunction with the Ninth Conference on Machine Translation (WMT) 2024. In this edition, participants were challenged to develop machine translation models for four distinct language pairs: English–Assamese, English-Mizo, English-Khasi, and English-Manipuri. The task utilized the enriched IndicNE-Corp1.0 dataset, which includes an extensive collection of parallel and monolingual corpora for northeastern Indic languages. The evaluation was conducted through a comprehensive suite of automatic metrics—BLEU, TER, RIBES, METEOR, and ChrF—supplemented by meticulous human assessment to measure the translation systems’ performance and accuracy. This initiative aims to drive advancements in low-resource machine translation and make a substantial contribution to the growing body of knowledge in this dynamic field.

Online sexism is a rising issue that threatens women’s safety, fosters hostile situations, and upholds social inequities. We describe a task SemEval-2023 Task 10 for creating English-language models that can precisely identify and categorize sexist content on internet forums and social platforms like Gab and Reddit as well to provide an explainability in order to address this problem. The problem is divided into three hierarchically organized subtasks: binary sexism detection, sexism by category, and sexism by fine-grained vector. The dataset consists of 20,000 labelled entries. For Task A, pertained models like Convolutional Neural Network (CNN) and Bidirectional Long Short-Term Memory (BiLSTM), which is called CNN-BiLSTM and Generative Pretrained Transformer 2 (GPT-2) models were used, as well as the GPT-2 model for Task B and C, and have provided experimental configurations. According to our findings, the GPT-2 model performs better than the CNN-BiLSTM model for Task A, while GPT-2 is highly accurate for Tasks B and C on the training, validation and testing splits of the training data provided in the task. Our proposed models allow researchers to create more precise and understandable models for identifying and categorizing sexist content in online forums, thereby empowering users and moderators.

In today’s digital culture, sexism and misogyny on online platforms have grown to be serious issues. To solve these problems, efficient automated sexist content detection and classification techniques must be created. In this study, we investigate the application of the GPT-2 model, a cutting-edge pre-trained language model, to the shared Exist 2023 job of sexism categorization. On the Exist 2023 dataset, we fine-tuned the GPT-2 model by adding adjustments like a classification head and weighted cross-entropy loss to tackle class imbalance. Our experimental findings show the GPT-2 model’s potential for precisely recognizing and classifying instances of sexism. Using the official assessment measure ICM (Information Contrast Measure), we assess our strategy while taking into account various evaluation modes, such as hard-hard, hard-soft, and soft-soft. The results show how well the GPT-2 model handles the problem of sexism categorization, assisting in the creation of automated techniques for fostering a safer and more welcoming online environment.

Hate speech continues to be a pressing concern in online social media (OSM) platforms, necessitating effective automated detection systems. In this paper, we propose a unified approach, encompassing both Task 1 & 4, to tackle the challenge of hate speech recognition within the HASOC 2023 framework. It addresses the complexities of multilingual OSM by employing cutting-edge Natural Language Processing (NLP) techniques and leveraging powerful language models put forward by team CNLP-NITS-PP. The key objective is optimising precision-recall trade-offs in hate speech detection, spanning English and Indo-Aryan languages. The empirical results demonstrate the effectiveness of our approach in isolating explicit signs of hate speech, emphasizing model efficiency, interpretability, and the importance of diverse linguistic nuances in creating safer online environments. This integrated work sets the stage for advancements in hate-span detection and underlines the significance of fostering responsible and inclusive online conversations across various language environments.

Hate speech poses significant challenges to maintaining healthy online conversations, and automated systems are crucial for its accurate detection and mitigation. In this paper, we (CNLP-NITS-PP) introduce ATLANTIS (Attentive Transformer-LSTM for Named Entity and Token Identification System), a robust model designed to address the pervasive issue of hate speech in online social media platforms. ATLANTIS focuses on hate span identification within sentences labeled as hate speech, framed as a sequence labeling task using BIO notation. Leveraging a Hate dataset enriched with Named Entity Recognition (NER) tags, ATLANTIS effectively identifies hate speech spans within the text by combining contextualized representations and sequential modeling. The empirical results showcase ATLANTIS’s effectiveness in isolating explicit signs of hate from a contextual backdrop, offering a promising solution for creating safer online environments. We achieve a macro F1 score of 0.488 on the public test set and 0.508 on the private test set. This work not only lays the foundation for future advancements in hate-span detection but also emphasizes the importance of model efficiency, interpretability, and expanded training data that encompass diverse linguistic nuances and evolving hate speech trends. Code is available at https://github.com/niyarrbarman/hasoc23.

In video surveillance schemes, the motion object detection plays a significant role. To subtract the object background, a segmentation technique based on feature extraction is utilized in which the change in the training rate makes an alteration in the background. Thereafter, the extracted features are trained by using the self-organizing map (SOM) network in which the weight parameters in the network is optimized with the help of artificial bee colony (ABC) optimization algorithm, so, the proposed methodology is named as HSOM-ABC technique. This methodology is carried out to perform the classification process in this research. Initially, the whole dataset is preprocessed with the help of grayscale conversion method which converts the original image into grayscale color. After this, fuzzy c-means clustering is applied to perform the segmentation process and this method divides the foreground and background parts efficiently. Then, feature extraction is done with the help of local binary pattern method which extract the relevant features from the segmented image. Finally, HSOM-ABC method is proposed to accurate classification process. Hence, the moving objects are identified by categorizing the background and foreground images. MatLab platform is chosen for the proposed work simulation and the performance is evaluated by means of different parameters and it is compared with new existing approaches. Experimental outcomes show that the proposed strategy achieves higher precision value than any other existing methods.