KhatianDoc: A Human-Verified Benchmark Diagnosing Multimodal LLM Failure on Bengali Legal Land Records

Land ownership in Bangladesh is recorded in Ana-Ganda-Kora-Kranti-Til, a base-16 positional fraction system with dedicated Unicode glyphs, no mainstream font, and no coverage in any OCR pipeline or tokenizer. The handwritten records that carry these fractions, RS Khatians, are the authoritative title record for millions of parcels and a frequent subject of civil litigation, yet no benchmark has asked whether a machine can read one. We introduce KhatianDoc, a four-task benchmark built from 107 real RS Khatian records from the Vumi (land) Office of Munshiganj, Bangladesh: symbol recognition, base-16-to-decimal conversion, structured field extraction, and legal document question answering over 1,634 QA pairs. Ground truth was transcribed by hand, verified by a land-law practitioner to full agreement, and anonymized through positional tokens that keep the referential distinctions multi-hop questions depend on. We evaluate six multimodal LLMs (8B to 72B+, open and closed) under a fixed zero-shot protocol. Five QA categories, 39.3% of our stratified set, return zero correct answers from every model; on the arithmetic task, every model that emits a number does worse than a constant-mean baseline, with exact- and near-match scores coinciding: decorrelation, not approximation. Auditing our own metrics surfaced two artifacts in opposite directions: we correct a refusal-scoring bug and report the fixed scores beside the originals, and flag an inflated metadata metric as an upper bound. KhatianDoc documents not a performance gap but the absence of a capability, with verified ground truth for future systems. Code and data, with a redacted image release, are publicly available.

DocTalkBN: A Novel Dataset of Expert Telemedicine Conversations in Bengali

Reliable medical conversational AI requires authentic expert--patient interaction data, yet such datasets remain scarce, especially for low-resource languages such as Bengali. We present DocTalkBN, a large-scale multimodal dataset of real-world expert telemedicine conversations in Bengali, collected from nationally broadcast telemedicine programs featuring board-certified physicians. DocTalkBN contains 557.63 hours of paired audio and text, 1,515 multi-turn patient calls, 10,274 host--doctor question--answer exchanges, totaling 1.7M tokens, spanning 26 medical specialties. Unlike prior resources derived from medical forums, written health content, or synthetic data, our dataset preserves the spontaneity, contextual richness, and spoken characteristics of authentic medical interactions in a low-resource setting. To support benchmark-driven research, we further construct three downstream tasks from the corpus, medical triage classification, advice safety evaluation, and medical named entity recognition, and benchmark a diverse set of large language models and encoder-based baselines. Our results show that DocTalkBN is a practically useful resource, particularly for clinically grounded reasoning tasks. We release this resource to facilitate future research on reliable medical NLP and safer, more culturally grounded healthcare systems for low-resource languages. Our source codes and dataset are publicly available at https://anonymous.4open.science/r/doctalk.

Register Shifts Break LLM Safety: A Bengali Benchmark with Culturally Grounded Harms

Bengali is the seventh-most-spoken language globally, yet LLM safety evaluation remains overwhelmingly English-centric. We introduce BanglaSafe, a benchmark of 879 Bengali prompts combining 309 natively authored prompts with 570 expert-reviewed prompts, spanning 17 culturally grounded harm categories and five prompting conditions that vary language, writing style, and authority framing. Evaluating 18 frontier LLMs, we find that over half of all responses are unsafe or partially unsafe (53.6%) while 14.7% contains strictly harmful content, and that the strongest observed effect is not the switch from English to Bengali but the choice of writing style within Bengali: the same harmful request phrased as a formal newspaper investigation succeeds 17 percentage points more often than the same request phrased as a casual message, with no adversarial engineering involved. We further show that existing safety classifiers struggle to reliably evaluate Bengali content, with even frontier models failing on nearly half of all cases.

When Less Is Enough: Context Selection and Prompting Strategies for Bengali News Headline Generation

Large language models (LLMs) have shown strong performance in text generation tasks, yet their effectiveness on headline generation remains sensitive to how input context is selected and presented. In this work, we investigate Bengali news headline generation as a document-level generation task that requires effective selection and presentation of salient contextual information from long-form articles. Using Gemini-2.0-Flash, Llama-3.3-70B, and GPT-4o, we systematically study the effects of context selection, prompting strategies, and in-context learning (i.e., few-shot) on the quality of headline generation. Our experiments show that providing the full article does not necessarily improve performance; instead, using selected lead paragraphs of the article can maintain, and in some cases improve, headline generation quality. We further compare Bengali Native Prompting (BNaP) and Cross-Lingual Prompting (XLP), and examine how each interacts with context-enriched prompt templates incorporating auxiliary contextual cues. Results demonstrate that prompting strategies substantially influence generation quality: XLP often yields stronger performance, particularly when combined with contextual enrichment, but its benefits are model-dependent. Additionally, few-shot prompting substantially improves Gemini, with most of the gain obtained from a single demonstration, whereas Llama shows limited benefit from additional examples. Overall, our findings highlight that effective Bengali news headline generation depends more on context relevance and prompt design than on increasing input length, offering practical insights for multilingual and low-resource LLM applications.

BengaliMCQ: Automatic Generation and Answer Prediction of Academic Multiple-Choice Questions in a Low-Resource Language

Traditional retrieval-augmented generation (RAG) frameworks process documents without attending to their hierarchical structure, leading to poor performance, especially in low-resource languages such as Bengali. To address this, we propose a structure-aware RAG framework that models Bengali textbooks as hierarchical graphs and uses a contrastively trained graph neural network to retrieve a small set of relevant passages. These passages provide focused context for a large language model, enabling topic-specific multiple-choice question (MCQ) generation and in-domain answer prediction. Experimental results demonstrate that our framework outperforms strong dense retrieval baselines across retrieval metrics, produces more relevant MCQs, and achieves superior answer prediction accuracy.

Where Does Retrieval Fail? Evaluating RAG Architectures for Agricultural Advisory

Retrieval quality in RAG systems is commonly reported as a single aggregate score, which can hide large differences across query types and language conditions. We study this problem in Bengali agricultural advisory, where farmer queries are often colloquial while official advisory documents use formal scientific terminology. We construct a test collection of 1,000 queries and 2,882 knowledge nodes extracted from 284 official Bangladeshi agricultural publications, and use it to evaluate five retrieval architectures and six embedding models under three controlled language conditions. The results show that no single retrieval method is consistently best. For native Bengali queries, BM25 is the strongest single retriever (R@10 = 0.506) while Hybrid RRF reaches the highest overall R@10 of 0.539. However, dense retrieval performance varies sharply by query type: R@10 is 0.093 on colloquial farmer queries and 0.970 on formal safety queries. Across language conditions, BM25 R@10 drops from 0.506 on Bengali queries to 0.004 when English queries are matched against the Bengali corpus, while dense retrieval falls only from 0.464 to 0.425. We also find that embedding task configuration and passage length can each change reported R@10 by a factor of seven, independent of architecture. These results show why low-resource RAG evaluation should report performance by language condition and query type rather than relying on aggregate scores alone. The dataset and evaluation scripts are available at https://huggingface.co/datasets/RaiyanKhaan/AgriTrust-RAG.

Structural Silence: When AI Infrastructure Fails Speakers of Underrepresented Languages

Artificial intelligence tools for education and language support are increasingly framed as scalable responses to access gaps in under-resourced communities. Yet the infrastructure underlying these tools, including training corpora, tokenization schemes, evaluation benchmarks, and deployment architectures, can systematically disadvantage speakers of underrepresented languages before a model is trained. This paper examines these structural barriers through Bengali, one of the world's most widely spoken languages, focusing on AI-assisted education in low-connectivity environments. We identify four interlocking failures: a severe web presence gap, with Bengali accounting for less than 0.5% of global web content despite representing nearly 4% of the global population; a 67:1 training-token deficit between English and Bengali in major multilingual corpora; a tokenization penalty associated with Bengali's alphasyllabary script that compounds the data deficit through higher token fertility; and connectivity exclusion, with individual internet penetration at 36.5% in rural areas compared with 71.4% in urban areas. These failures reflect longstanding resource-allocation decisions, institutional priorities, and design defaults that did not center underrepresented languages in mainstream AI development. We argue that dataset scarcity should be understood as a structural barrier rather than an isolated technical limitation, and that offline-first design should be treated as an equity-oriented infrastructure strategy. We conclude with directions for linguistics and AI research aimed at reducing these structural inequalities.

BanglaWild: An In-the-Wild Bengali Scene Text Recognition Benchmark for OCR and Vision-Language Models

In-the-wild Bengali scene text recognition is largely unmeasured: existing resources target handwritten documents or constrained sign-board parsing, report only aggregate edit-distance metrics, and evaluate either conventional OCR or VLMs, never both on the same in-the-wild data. To address this gap, we introduce BANGLAWILD, a benchmark of 2,535 Bengali scene text images, each paired with a verbatim gold transcription, two categorical axes, four diagnostic attributes, and an orthographically standard form where the in-image text deviates from canonical spelling. We evaluate fifteen VLMs and three conventional OCR systems under three prompting strategies, fine-tune 6 open-source models with LoRA, and complement edit-distance metrics with an LLM-as-a-Judge evaluation. Our results reveal a persistent gap in which larger models within the same family do not outperform smaller ones. Our fifteen-class error taxonomy shows that visual mis-recognition accounts for ~60% of errors in the strongest systems, while conjunct-related errors contribute under 2%, challenging a long-standing assumption in Bengali OCR research; the same visual dominant profile also holds across architectures, including the one conventional baseline that reads Bengali reliably. Prompt language mainly affects cross-script drift and LoRA reduces catastrophic failures in weak models without lifting the ceiling on already competent ones. Code and data will be publicly released.

Two-Stage Bengali Sentiment Classification: Domain Adaptation Through Continual Learning and Parameter-Efficient Fine-Tuning

Understanding sentiment in low-resource languages remains a key challenge for Natural Language Processing (NLP), particularly when domain-specific data is scarce. In this work, we present SentiBanglaBERT, a two-stage Bengali sentiment classification framework combining domain-adaptive continual pretraining and parameter-efficient fine-tuning. The approach enables contextual adaptation to news-style data while remaining computationally efficient through Low-Rank Adaptation (LoRA). Beyond performance, SentiBanglaBERT integrates SHAP-based interpretability, offering linguistic insights into how Bengali morphological cues, such as negation suffixes and aspectual markers, influence sentiment predictions. Experiments demonstrate stable performance comparable to strong baselines while providing greater transparency and interpretive depth. This framework highlights the potential of domain-adaptive continual learning as a foundation for interpretable, resource-efficient NLP in morphologically rich, underrepresented languages.

Token-Region Guided Cross-Attention Fusion for Multimodal Affect Interpretation

Automated analysis of multimodal content on social networks has become a critical task for understanding public sentiment and information diffusion in the digital age. However, classifying internet memes remains computationally challenging due to the intricate interplay between visual cues and embedded, often stylized, text, particularly in low-resource languages like Bengali Language. This paper addresses the detection of political intent in Bengali memes by introducing Multimodal Cross-Attention Fusion framework. We first leverage a Vision-Language Model to extract high-fidelity OCR text from noisy meme images. Subsequently, we encode visual and textual features and synthesize them through a cross-modal multi-head attention mechanism that aligns semantic tokens with visual regions. We also investigate the integration of a domain-specific political lexicon as a knowledge prior. Experimental evaluation on the PoliMemeDecode1 dataset shows that our attention-based fusion significantly outperforms unimodal baselines and standard concatenation methods, achieving a state-of-the-art Macro-F1 of approximately 0.94. Interpretability analyzes further confirm that the model effectively learns to ground textual semantics in visual evidence.

PatiGonit22K: A Comprehensive Dataset for Solving Complex Bengali MWPs

Mathematical Word Problems (MWPs) are an important benchmark for evaluating natural language understanding and quantitative reasoning. Despite recent progress in high resource languages, Bengali remains underexplored due to the limited availability of large scale annotated datasets. In this work, we introduce PatiGonit22K, an expanded Bengali MWP dataset containing 22,441 problems, developed by extending the original PatiGonit dataset with a substantially larger collection of complex mathematical problems. The dataset includes both simple and multi operation equations, providing a balanced benchmark for evaluating mathematical reasoning across different difficulty levels. Each problem is carefully translated, annotated, culturally adapted, and verified to ensure linguistic consistency and mathematical correctness. By increasing both the scale and complexity of Bengali MWPs, PatiGonit22K provides a more comprehensive resource for future research on mathematical reasoning and educational NLP applications in low resource languages.

BanClickThumb: A Multimodal Dataset and Transformer Fusion Benchmarks for Clickbait Detection in Bengali YouTube Videos

Clickbait, where video titles and thumbnails exaggerate or misrepresent content, reduces user trust, wastes attention, and promotes misinformation on video-sharing platforms. Detecting Bengali clickbait remains challenging because publicly available multimodal datasets are limited. To address this gap, we introduce BanClickThumb, a curated dataset of 7,147 Bengali YouTube thumbnail-title pairs from five content domains, annotated by ten annotators with high agreement (Cohen's Kappa: 0.83-0.93). Using this dataset, we benchmark text-only, image-only, and multimodal approaches. Among unimodal models, BanClickTextFormer (XLM-RoBERTa) achieves 0.82 accuracy, while BanClickImageFormer (SwiftFormer) reaches 0.68. Our proposed multimodal model, BanClickFusionFormer, combines ViT and XLM-RoBERTa through intermediate fusion and achieves the best accuracy of 0.84. Error analysis shows that dense thumbnail text, figurative language, and culturally specific slang remain challenging. Our findings demonstrate the effectiveness of multimodal fusion for Bengali clickbait detection and provide a publicly available benchmark to support future research on low-resource multimodal content analysis.

GSM-Plus-BN: A Perturbation-Based Benchmark for Bangla Mathematical Reasoning in Large Language Models

The evaluation of mathematical reasoning in large language models (LLMs) has predominantly focused on high-resource languages like English. This has created a significant barrier to the equitable development and deployment of AI in linguistically diverse regions such as Bangladesh, where over 230 million people speak Bengali. Despite this global significance, there has been minimal prior work on mathematical reasoning in Bengali and no existing research that systematically benchmarks a perturbated Bengali mathematical dataset, leaving a critical void in assessing model robustness and true comprehension beyond pattern recognition. This study addresses this gap by introducing GSM-Plus-BN, a novel perturbated Bengali mathematical dataset derived from the English GSM-Plus benchmark and verified by human translators. We evaluate six open-source LLMs Qwen3-32B, Llama-3.1-8B-Instant, Llama-3.3-70B-Versatile, Llama-4-Scout-17B-16E-Instruct, GPT-OSS-120B, and GPT-OSS-20B using a benchmark of 9,000 evaluation samples comprising 1,000 seed questions and 8,000 perturbed variants under both Standard Prompting and Chain-of-Thought (CoT) Prompting. Experimental results show that GPT-OSS-20B achieves the highest seed question accuracy of 96.08% under Standard Prompting, while larger models such as Llama-3.3-70B and GPT-OSS-120B demonstrate superior robustness across perturbation types. Furthermore, CoT prompting substantially improves reasoning for most models compared to Standard Prompting, yet a notable performance gap persists across all models relative to their English benchmarks, underscoring the inherent difficulty of perturbed Bengali text. This research makes a foundational contribution by providing GSM-PLUS-BN as a new resource and baseline for future Bengali mathematical reasoning research.

Tokenizer Transplantation: Mitigating Autoregressive Collapse in Edge-Efficient Bengali ASR

Lightweight speech recognition models are critical for edge deployment, yet highly optimized architectures like Moonshine often fail on morphologically rich, non-Latin languages such as Bengali. This study identifies the root cause of this failure as the model's English-centric byte-level tokenizer, which fragments Bengali words into high-fertility byte chains and triggers catastrophic autoregressive collapse during inference. To resolve this, a novel vocabulary transplantation pipeline is proposed to replace the decoder vocabulary with the native-script BanglaBERT WordPiece vocabulary and resize the corresponding token embedding matrix. Experimental results demonstrate a reduction in token fertility from 9.16 to 1.30. By decreasing autoregressive sequence length by 85.8%, decoding instability is entirely mitigated. When evaluated on the 882-hour Lipi-Ghor dataset, the modified architecture achieves a competitive 21.54% Word Error Rate (WER) and a Real-Time Factor (RTF) of 0.0053. Ultimately, this research provides a scalable, reproducible blueprint for cross-script adaptation of compact ASR models without the need for resource-intensive pre-training.

BanglaMemeEvidence: A Multimodal Benchmark Dataset for Explanatory Evidence Detection in Bengali Memes

Memes have become influential communication tools on social media, combining viral visuals with concise messaging to convey impactful ideas. While substantial research has examined the affective dimensions of memes, key challenges such as detecting harmful content, identifying cyberbullying, and performing accurate sentiment analysis remain critical, largely due to the need for deeper contextual understanding. In this paper, we introduce MemeEvidenceDetect, a hybrid task aimed at analyzing a meme and its contextual information to identify specific sentences that explain or elucidate its meaning and humor. To support this task, we present BanglaMemeEvidence, a curated dataset of 2,917 Bengali memes, emphasizing its significance as a resource for the Bangla language. Each meme is annotated with natural language explanations, including Meme OCR, Meme Context, and Evidence Sentences, alongside relevance scores that reflect the relationship between a meme and its corresponding annotations. To address the gap in dynamically inferring a meme's context, we propose BengaliMemeEvidenceNet, a hybrid multimodal framework that integrates textual and visual features for comprehensive meme representation. Our experiments demonstrate the effectiveness of BengaliMemeEvidenceNet, achieving an F1 score of 0.74. To the best of our knowledge, this is the first study to focus on evidence detection in Bengali memes, marking a notable step forward in the analysis of memes in low-resource languages.

Building and Evaluating a Synthetic Bengali Speech Resource for Telecom Customer Care

Speech systems used in customer-facing applications often require domain-specific language coverage. We present a synthetic Bengali speech dataset for telecom customer-care scenarios. The dataset contains 10,000 audio-text pairs, approximately 26.82 hours of 24 kHz speech, and predefined train, validation, and test splits of 9,000, 500, and 500 examples. It is publicly released on Hugging Face under the CC-BY-4.0 license. The speech was generated with OmniVoice in voice-cloning mode using a real female reference recording and transcript, with bfloat16 precision, 16 diffusion sampling steps, and a speaking-rate control value of 1.0. Along with the original Bengali text, the dataset provides a normalized transcript field designed for ASR/STT training and evaluation. We report an automatic intelligibility check over all 10,000 samples using a domain-adapted Whisper ASR model fine-tuned from bengaliAI/tugstugi_bengaliai-regional-asr_whisper-medium, along with a manual listening check on selected samples. The evaluation gives an average WER of 2.54%, an average CER of 0.59%, and median WER and CER values of 0.00%. These results suggest strong text-audio consistency under the selected automatic evaluation pipeline, while the paper also discusses the limitations of synthetic speech and STT-based evaluation.

BenSyc: Benchmarking Conversational Sycophancy and Human Alignment in LLMs for Bengali Contexts

Large language models (LLMs) increasingly participate in emotionally sensitive social conversations, where responses may shift from balanced support toward excessive validation or escalatory alignment. Existing sycophancy research primarily focuses on factual agreement and instruction-following settings, leaving culturally grounded conversational sycophancy underexplored. We introduce BenSyc, the first benchmark for studying conversational sycophancy in Bengali social contexts. Starting from 11,840 Reddit posts and 170k comments collected from communities across Bangladesh and West Bengal, we construct a human-validated benchmark with binary labels and a fine-grained five-level taxonomy spanning Invalidation, Neutral, Support, Validation, and Escalation. We evaluate more than 15 open and proprietary LLMs on conversational alignment classification and response generation tasks. Results show that distinguishing empathetic support from reinforcement-oriented validation remains challenging even for frontier instruction-tuned models: the best system achieves only 61.8 Macro-F1 on binary detection and 61.7 Macro-F1 on five-class classification. In generation settings, several models frequently produce strongly validating or escalatory responses in emotionally charged situations. Our findings highlight substantial variation across model families and conversational behaviors, underscoring the importance of culturally grounded multilingual benchmarks for evaluating socially aligned conversational AI systems.