Realignment is a promising strategy to improve cross-lingual transfer in multilingual language models. However, empirical results are mixed and often unreliable, particularly for typologically distant or low-resource languages (LRLs) compared to English. Moreover, word realignment tools often rely on high-quality parallel data, which can be scarce or noisy for many LRLs. In this work, we conduct an extensive empirical study to investigate whether realignment truly benefits from using all available languages, or if strategically selected subsets can offer comparable or even improved cross-lingual transfer, and study the impact on LRLs. Our controlled experiments show that realignment can be particularly effective for LRLs and that using carefully selected, linguistically diverse subsets can match full multilingual alignment, and even outperform it for unseen LRLs. This indicates that effective realignment does not require exhaustive language coverage and can reduce data collection overhead, while remaining both efficient and robust when guided by informed language selection.

In this paper, we combine two-step knowledge distillation, structured pruning, truncation, and vocabulary trimming for extremely compressing multilingual encoder-only language models for low-resource languages. Our novel approach systematically combines existing techniques and takes them to the extreme, reducing layer depth, feed-forward hidden size, and intermediate layer embedding size to create significantly smaller monolingual models while retaining essential language-specific knowledge. We achieve compression rates of up to 92% while maintaining competitive performance, with average drops of 2-10% for moderate compression and 8-13% at maximum compression in four downstream tasks, including sentiment analysis, topic classification, named entity recognition, and part-of-speech tagging, across three low-resource languages. Notably, the performance degradation correlates with the amount of language-specific data in the teacher model, with larger datasets resulting in smaller performance losses. Additionally, we conduct ablation studies to identify the best practices for multilingual model compression using these techniques.

This is particularly true for language tasks that are culture-dependent.

Hope speech language that fosters encouragement and optimism plays a vital role in promoting positive discourse online. However, its detection remains challenging, especially in multilingual and low-resource settings. This paper presents a multilingual framework for hope speech detection using an active learning approach and transformer-based models, including mBERT and XLM-RoBERTa. Experiments were conducted on datasets in English, Spanish, German, and Urdu, including benchmark test sets from recent shared tasks. Our results show that transformer models significantly outperform traditional baselines, with XLM-RoBERTa achieving the highest overall accuracy. Furthermore, our active learning strategy maintained strong performance even with small annotated datasets. This study highlights the effectiveness of combining multilingual transformers with data-efficient training strategies for hope speech detection.

The July Revolution in Bangladesh marked a significant student-led mass uprising, uniting people across the nation to demand justice, accountability, and systemic reform. Social media platforms played a pivotal role in amplifying public sentiment and shaping discourse during this historic mass uprising. In this study, we present a hybrid transformer-based sentiment analysis framework to decode public opinion expressed in social media comments during and after the revolution. We used a brand new dataset of 4,200 Bangla comments collected from social media. The framework employs advanced transformer-based feature extraction techniques, including BanglaBERT, mBERT, XLM-RoBERTa, and the proposed hybrid XMB-BERT, to capture nuanced patterns in textual data. Principle Component Analysis (PCA) were utilized for dimensionality reduction to enhance computational efficiency. We explored eleven traditional and advanced machine learning classifiers for identifying sentiments. The proposed hybrid XMB-BERT with the voting classifier achieved an exceptional accuracy of 83.7% and outperform other model classifier combinations. This study underscores the potential of machine learning techniques to analyze social sentiment in low-resource languages like Bangla.

In natural language processing, multilingual models like mBERT and XLM-RoBERTa promise broad coverage but often struggle with languages that share a script yet differ in orthographic norms and cultural context. This issue is especially notable in Arabic-script languages such as Kurdish Sorani, Arabic, Persian, and Urdu. We introduce the Arabic Script RoBERTa (AS-RoBERTa) family: four RoBERTa-based models, each pre-trained on a large corpus tailored to its specific language. By focusing pre-training on language-specific script features and statistics, our models capture patterns overlooked by general-purpose models. When fine-tuned on classification tasks, AS-RoBERTa variants outperform mBERT and XLM-RoBERTa by 2 to 5 percentage points. An ablation study confirms that script-focused pre-training is central to these gains. Error analysis using confusion matrices shows how shared script traits and domain-specific content affect performance. Our results highlight the value of script-aware specialization for languages using the Arabic script and support further work on pre-training strategies rooted in script and language specificity.

--The rapid expansion of social media platforms has significantly increased the dissemination of forged content and misinformation, making the detection of fake news a critical area of research. Although fact-checking efforts predominantly focus on English-language news, there is a noticeable gap in resources and strategies to detect news in regional languages, such as Urdu. Advanced Fake News Detection (FND) techniques rely heavily on large, accurately labeled datasets. However, FND in under-resourced languages like Urdu faces substantial challenges due to the scarcity of extensive corpora and the lack of validated lexical resources. Current Urdu fake news datasets are often domain-specific and inaccessible to the public. They also lack human verification, relying mainly on unverified English-to-Urdu translations, which compromises their reliability in practical applications. This study highlights the necessity of developing reliable, expert-verified, and domain-independent Urdu-enhanced FND datasets to improve fake news detection in Urdu and other resource-constrained languages. This paper presents the first benchmark large FND dataset for Urdu news, which is publicly available for validation and deep analysis. We also evaluate this dataset using multiple state-of-the-art pre-trained large language models (LLMs), such as XLNet, mBERT, XLM-RoBERT a, RoBERT a, DistilBERT, and DeBERT a. Additionally, we propose a unified LLM model that outperforms the others with different embedding and feature extraction techniques. The performance of these models is compared based on accuracy, F1 score, precision, recall, and human judgment for vetting the sample results of news. The proposed model outperforms advanced machine learning and deep learning models previously used in the literature for fake news detection.

While transformer-based models achieve strong performance on text classification, we explore whether masking input tokens can further enhance their effectiveness. We propose token masking regularization, a simple yet theoretically motivated method that randomly replaces input tokens with a special [MASK] token at probability p. This introduces stochastic perturbations during training, leading to implicit gradient averaging that encourages the model to capture deeper inter-token dependencies. Experiments on language identification and sentiment analysis -- across diverse models (mBERT, Qwen2.5-0.5B, TinyLlama-1.1B) -- show consistent improvements over standard regularization techniques. We identify task-specific optimal masking rates, with p = 0.1 as a strong general default. We attribute the gains to two key effects: (1) input perturbation reduces overfitting, and (2) gradient-level smoothing acts as implicit ensembling.

Code-switching, or alternating between languages within a single conversation, presents challenges for multilingual language models on NLP tasks. This research investigates if pre-training Multilingual BERT (mBERT) on code-switched datasets improves the model's performance on critical NLP tasks such as part of speech tagging, sentiment analysis, named entity recognition, and language identification. We use a dataset of Spanglish tweets for pre-training and evaluate the pre-trained model against a baseline model. Our findings show that our pre-trained mBERT model outperforms or matches the baseline model in the given tasks, with the most significant improvements seen for parts of speech tagging. Additionally, our latent analysis uncovers more homogenous English and Spanish embeddings for language identification tasks, providing insights for future modeling work. This research highlights potential for adapting multilingual LMs for code-switched input data in order for advanced utility in globalized and multilingual contexts. Future work includes extending experiments to other language pairs, incorporating multiform data, and exploring methods for better understanding context-dependent code-switches.