Swakkhar Shatabda

Mahbub E Sobhani, Md. Faiyaz Abdullah Sayeedi, Mohammad Nehad Alam et al.

Diagram-grounded geometry problem solving is a critical benchmark for multimodal large language models (MLLMs), yet the benefits of multi-agent design over single-agent remain unclear. We systematically compare single-agent and multi-agent pipelines on four visual math benchmarks: Geometry3K, MathVerse, OlympiadBench, and We-Math. For open-source models, multi-agent consistently improves performance. For example, Qwen-2.5-VL (7B) gains +6.8 points and Qwen-2.5-VL (32B) gains +3.3 on Geometry3K, and both Qwen-2.5-VL variants see further gains on OlympiadBench and We-Math. In contrast, the closed-source Gemini-2.0-Flash generally performs better in single-agent mode on classic benchmarks, while multi-agent yields only modest improvements on the newer We-Math dataset. These findings show that multi-agent pipelines provide clear benefits for open-source models and can assist strong proprietary systems on newer, less familiar benchmarks, but agentic decomposition is not universally optimal. All code, data, and reasoning files are available at https://github.com/faiyazabdullah/Interpreter-Solver

Nurul Labib Sayeedi, Md. Faiyaz Abdullah Sayeedi, Khushnur Binte Jahangir et al.

Large Language Models (LLMs) show impressive performance on many NLP benchmarks, yet their ability to reason in figurative, culturally grounded, and low-resource settings remains underexplored. We address this gap for Bangla by introducing BanglaRiddleEval, a benchmark of 1,244 traditional Bangla riddles instantiated across four tasks (4,976 riddle-task artifacts in total). Using an LLM-based pipeline, we generate Chain-of-Thought explanations, semantically coherent distractors, and fine-grained ambiguity annotations, and evaluate a diverse suite of open-source and closed-source models under different prompting strategies. Models achieve moderate semantic overlap on generative QA but low correctness, MCQ accuracy peaks at only about 56% versus an 83% human baseline, and ambiguity resolution ranges from roughly 26% to 68%, with high-quality explanations confined to the strongest models. These results show that current LLMs capture some cues needed for Bangla riddle reasoning but remain far from human-level performance, establishing BanglaRiddleEval as a challenging new benchmark for low-resource figurative reasoning. All data, code, and evaluation scripts are available on GitHub: https://github.com/Labib1610/BanglaRiddleEval.

Md. Rayhan Ahmed, Adnan Ferdous Ashrafi, Raihan Uddin Ahmed et al.

Accurate segmentation of the region of interest in medical images can provide an essential pathway for devising effective treatment plans for life-threatening diseases. It is still challenging for U-Net, and its state-of-the-art variants, such as CE-Net and DoubleU-Net, to effectively model the higher-level output feature maps of the convolutional units of the network mostly due to the presence of various scales of the region of interest, intricacy of context environments, ambiguous boundaries, and multiformity of textures in medical images. In this paper, we exploit multi-contextual features and several attention strategies to increase networks' ability to model discriminative feature representation for more accurate medical image segmentation, and we present a novel dual U-Net-based architecture named DoubleU-NetPlus. The DoubleU-NetPlus incorporates several architectural modifications. In particular, we integrate EfficientNetB7 as the feature encoder module, a newly designed multi-kernel residual convolution module, and an adaptive feature re-calibrating attention-based atrous spatial pyramid pooling module to progressively and precisely accumulate discriminative multi-scale high-level contextual feature maps and emphasize the salient regions. In addition, we introduce a novel triple attention gate module and a hybrid triple attention module to encourage selective modeling of relevant medical image features. Moreover, to mitigate the gradient vanishing issue and incorporate high-resolution features with deeper spatial details, the standard convolution operation is replaced with the attention-guided residual convolution operations, ...

Mehedi Hasan Bijoy, Nahid Hossain, Salekul Islam et al.

Spelling error correction is the task of identifying and rectifying misspelled words in texts. It is a potential and active research topic in Natural Language Processing because of numerous applications in human language understanding. The phonetically or visually similar yet semantically distinct characters make it an arduous task in any language. Earlier efforts on spelling error correction in Bangla and resource-scarce Indic languages focused on rule-based, statistical, and machine learning-based methods which we found rather inefficient. In particular, machine learning-based approaches, which exhibit superior performance to rule-based and statistical methods, are ineffective as they correct each character regardless of its appropriateness. In this paper, we propose a novel detector-purificator-corrector framework, DPCSpell based on denoising transformers by addressing previous issues. In addition to that, we present a method for large-scale corpus creation from scratch which in turn resolves the resource limitation problem of any left-to-right scripted language. The empirical outcomes demonstrate the effectiveness of our approach, which outperforms previous state-of-the-art methods by attaining an exact match (EM) score of 94.78%, a precision score of 0.9487, a recall score of 0.9478, an f1 score of 0.948, an f0.5 score of 0.9483, and a modified accuracy (MA) score of 95.16% for Bangla spelling error correction. The models and corpus are publicly available at https://tinyurl.com/DPCSpell.

Al Mohimanul Islam, Sadia Shakiba Bhuiyan, Mysun Mashira et al.

Segmenting polyps in colonoscopy images is essential for the early identification and diagnosis of colorectal cancer, a significant cause of worldwide cancer deaths. Prior deep learning based models such as Attention based variation, UNet variations and Transformer-derived networks have had notable success in capturing intricate features and complex polyp shapes. In this study, we have introduced the DeepLabv3++ model which is an enhanced version of the DeepLabv3+ architecture. It is designed to improve the precision and robustness of polyp segmentation in colonoscopy images. We have utilized The proposed model incorporates diverse separable convolutional layers and attention mechanisms within the MSPP block, enhancing its capacity to capture multi-scale and directional features. Additionally, the redesigned decoder further transforms the extracted features from the encoder into a more meaningful segmentation map. Our model was evaluated on three public datasets (CVC-ColonDB, CVC-ClinicDB, Kvasir-SEG) achieving Dice coefficient scores of 96.20%, 96.54%, and 96.08%, respectively. The experimental analysis shows that DeepLabV3++ outperforms several state-of-the-art models in polyp segmentation tasks. Furthermore, compared to the baseline DeepLabV3+ model, our DeepLabV3++ with its MSPP module and redesigned decoder architecture, significantly reduced segmentation errors (e.g., false positives/negatives) across small, medium, and large polyps. This improvement in polyp delineation is crucial for accurate clinical decision-making in colonoscopy.

Mohammad Shamim Ahsan, Salekul Islam, Swakkhar Shatabda

The widespread adoption of the Internet of Things (IoT) has raised a new challenge for developers since it is prone to known and unknown cyberattacks due to its heterogeneity, flexibility, and close connectivity. To defend against such security breaches, researchers have focused on building sophisticated intrusion detection systems (IDSs) using machine learning (ML) techniques. Although these algorithms notably improve detection performance, they require excessive computing power and resources, which are crucial issues in IoT networks considering the recent trends of decentralized data processing and computing systems. Consequently, many optimization techniques have been incorporated with these ML models. Specifically, a special category of optimizer adopted from the behavior of living creatures and different aspects of natural phenomena, known as metaheuristic algorithms, has been a central focus in recent years and brought about remarkable results. Considering this vital significance, we present a comprehensive and systematic review of various applications of metaheuristics algorithms in developing a machine learning-based IDS, especially for IoT. A significant contribution of this study is the discovery of hidden correlations between these optimization techniques and machine learning models integrated with state-of-the-art IoT-IDSs. In addition, the effectiveness of these metaheuristic algorithms in different applications, such as feature selection, parameter or hyperparameter tuning, and hybrid usages are separately analyzed. Moreover, a taxonomy of existing IoT-IDSs is proposed. Furthermore, we investigate several critical issues related to such integration. Our extensive exploration ends with a discussion of promising optimization algorithms and technologies that can enhance the efficiency of IoT-IDSs.

Nurul Labib Sayeedi, Md. Faiyaz Abdullah Sayeedi, Shubhashis Roy Dipta et al.

Bangla culture is richly expressed through region, dialect, history, food, politics, media, and everyday visual life, yet it remains underrepresented in multimodal evaluation. To address this gap, we introduce BanglaVerse, a culturally grounded benchmark for evaluating multilingual vision-language models (VLMs) on Bengali culture across historically linked languages and regional dialects. Built from 1,152 manually curated images across nine domains, the benchmark supports visual question answering and captioning, and is expanded into four languages and five Bangla dialects, yielding ~32.3K artifacts. Our experiments show that evaluating only standard Bangla overestimates true model capability: performance drops under dialectal variation, especially for caption generation, while historically linked languages such as Hindi and Urdu retain some cultural meaning but remain weaker for structured reasoning. Across domains, the main bottleneck is missing cultural knowledge rather than visual grounding alone, with knowledge-intensive categories. These findings position BanglaVerse as a more realistic test bed for measuring culturally grounded multimodal understanding under linguistic variation.

Niful Islam, Md. Rayhan Ahmed, Nur Mohammad Fahad et al.

Remote sensing imagery plays a crucial role in many applications and requires accurate computerized classification techniques. Reliable classification is essential for transforming raw imagery into structured and usable information. While Convolutional Neural Networks (CNNs) are mostly used for image classification, they excel at local feature extraction, but struggle to capture global contextual information. Vision Transformers (ViTs) address this limitation through self attention mechanisms that model long-range dependencies. Integrating CNNs and ViTs, therefore, leads to better performance than standalone architectures. However, the use of additional CNN and ViT components does not lead to further performance improvement and instead introduces a bottleneck caused by redundant feature representations. In this research, we propose a fusion model that combines the strengths of CNNs and ViTs for remote sensing image classification. To overcome the performance bottleneck, the proposed approach trains four independent fusion models that integrate CNN and ViT backbones and combine their outputs at the final prediction stage through ensembling. The proposed method achieves accuracy rates of 98.10 percent, 94.46 percent, and 95.45 percent on the UC Merced, RSSCN7, and MSRSI datasets, respectively. These results outperform competing architectures and highlight the effectiveness of the proposed solution, particularly due to its efficient use of computational resources during training.

Mohammad Nazmush Shamael, Sabila Nawshin, Swakkhar Shatabda et al.

This work presents the BanglishRev Dataset, the largest e-commerce product review dataset to date for reviews written in Bengali, English, a mixture of both and Banglish, Bengali words written with English alphabets. The dataset comprises of 1.74 million written reviews from 3.2 million ratings information collected from a total of 128k products being sold in online e-commerce platforms targeting the Bengali population. It includes an extensive array of related metadata for each of the reviews including the rating given by the reviewer, date the review was posted and date of purchase, number of likes, dislikes, response from the seller, images associated with the review etc. With sentiment analysis being the most prominent usage of review datasets, experimentation with a binary sentiment analysis model with the review rating serving as an indicator of positive or negative sentiment was conducted to evaluate the effectiveness of the large amount of data presented in BanglishRev for sentiment analysis tasks. A BanglishBERT model is trained on the data from BanglishRev with reviews being considered labeled positive if the rating is greater than 3 and negative if the rating is less than or equal to 3. The model is evaluated by being testing against a previously published manually annotated dataset for e-commerce reviews written in a mixture of Bangla, English and Banglish. The experimental model achieved an exceptional accuracy of 94\% and F1 score of 0.94, demonstrating the dataset's efficacy for sentiment analysis. Some of the intriguing patterns and observations seen within the dataset and future research directions where the dataset can be utilized is also discussed and explored. The dataset can be accessed through https://huggingface.co/datasets/BanglishRev/bangla-english-and-code-mixed-ecommerce-review-dataset.

Soumik Deb Niloy, Md. Fahmid-Ul-Alam Juboraj, Swakkhar Shatabda

Pro-inflammatory peptides (PIPs) play critical roles in immune signaling and inflammation but are difficult to identify experimentally due to costly and time-consuming assays. To address this challenge, we present KEMP-PIP, a hybrid machine learning framework that integrates deep protein embeddings with handcrafted descriptors for robust PIP prediction. Our approach combines contextual embeddings from pretrained ESM protein language models with multi-scale k-mer frequencies, physicochemical descriptors, and modlAMP sequence features. Feature pruning and class-weighted logistic regression manage high dimensionality and class imbalance, while ensemble averaging with an optimized decision threshold enhances the sensitivity--specificity balance. Through systematic ablation studies, we demonstrate that integrating complementary feature sets consistently improves predictive performance. On the standard benchmark dataset, KEMP-PIP achieves an MCC of 0.505, accuracy of 0.752, and AUC of 0.762, outperforming ProIn-fuse, MultiFeatVotPIP, and StackPIP. Relative to StackPIP, these results represent improvements of 9.5% in MCC and 4.8% in both accuracy and AUC. The KEMP-PIP web server is freely available at https://nilsparrow1920-kemp-pip.hf.space/ and the full implementation at https://github.com/S18-Niloy/KEMP-PIP.

Mahbub E Sobhani, Md. Faiyaz Abdullah Sayeedi, Tasnim Mohiuddin et al.

Mathematical reasoning remains one of the most challenging domains for large language models (LLMs), requiring not only linguistic understanding but also structured logical deduction and numerical precision. While recent LLMs demonstrate strong general-purpose reasoning abilities, their mathematical competence across diverse languages remains underexplored. Existing benchmarks primarily focus on English or a narrow subset of high-resource languages, leaving significant gaps in assessing multilingual and cross-lingual mathematical reasoning. To address this, we introduce MathMist, a parallel multilingual benchmark for mathematical problem solving and reasoning. MathMist encompasses over 21K aligned question-answer pairs across seven languages, representing a balanced coverage of high-, medium-, and low-resource linguistic settings. The dataset captures linguistic variety, multiple types of problem settings, and solution synthesizing capabilities. We systematically evaluate a diverse suite of models, including open-source small and medium LLMs, proprietary systems, and multilingual-reasoning-focused models, under zero-shot, chain-of-thought (CoT), and code-switched reasoning paradigms. Our results reveal persistent deficiencies in LLMs' ability to perform consistent and interpretable mathematical reasoning across languages, with pronounced degradation in low-resource settings. All the codes and data are available at GitHub: https://github.com/mahbubhimel/MathMist

Md. Faiyaz Abdullah Sayeedi, Md. Mahbub Alam, Subhey Sadi Rahman et al.

The rise of Large Language Models (LLMs) has redefined Machine Translation (MT), enabling context-aware and fluent translations across hundreds of languages and textual domains. Despite their remarkable capabilities, LLMs often exhibit uneven performance across language families and specialized domains. Moreover, recent evidence reveals that these models can encode and amplify different biases present in their training data, posing serious concerns for fairness, especially in low-resource languages. To address these gaps, we introduce Translation Tangles, a unified framework and dataset for evaluating the translation quality and fairness of open-source LLMs. Our approach benchmarks 24 bidirectional language pairs across multiple domains using different metrics. We further propose a hybrid bias detection pipeline that integrates rule-based heuristics, semantic similarity filtering, and LLM-based validation. We also introduce a high-quality, bias-annotated dataset based on human evaluations of 1,439 translation-reference pairs. The code and dataset are accessible on GitHub: https://github.com/faiyazabdullah/TranslationTangles

Nirjhor Datta, Swakkhar Shatabda, M Sohel Rahman

Large pre-trained DNA language models such as DNABERT-2, Nucleotide Transformer, and HyenaDNA have demonstrated strong performance on various genomic benchmarks. However, most applications rely on expensive fine-tuning, which works best when the training and test data share a similar distribution. In this work, we investigate whether task-specific fine-tuning is always necessary. We show that simple embedding-based pipelines that extract fixed representations from these models and feed them into lightweight classifiers can achieve competitive performance. In evaluation settings with different data distributions, embedding-based methods often outperform fine-tuning while reducing inference time by 10x to 20x. Our results suggest that embedding extraction is not only a strong baseline but also a more generalizable and efficient alternative to fine-tuning, especially for deployment in diverse or unseen genomic contexts. For example, in enhancer classification, HyenaDNA embeddings combined with zCurve achieve 0.68 accuracy (vs. 0.58 for fine-tuning), with an 88% reduction in inference time and over 8x lower carbon emissions (0.02 kg vs. 0.17 kg CO2). In non-TATA promoter classification, DNABERT-2 embeddings with zCurve or GC content reach 0.85 accuracy (vs. 0.89 with fine-tuning) with a 22x lower carbon footprint (0.02 kg vs. 0.44 kg CO2). These results show that embedding-based pipelines offer over 10x better carbon efficiency while maintaining strong predictive performance. The code is available here: https://github.com/NIRJHOR-DATTA/EMBEDDING-IS-ALMOST-ALL-YOU-NEED.

Md. Adnanul Islam, Md. Faiyaz Abdullah Sayeedi, Md. Asaduzzaman Shuvo et al.

Mosquito-borne diseases pose a major global health risk, requiring early detection and proactive control of breeding sites to prevent outbreaks. In this paper, we present VisText-Mosquito, a multimodal dataset that integrates visual and textual data to support automated detection, segmentation, and reasoning for mosquito breeding site analysis. The dataset includes 1,828 annotated images for object detection, 142 images for water surface segmentation, and natural language reasoning texts linked to each image. The YOLOv9s model achieves the highest precision of 0.92926 and mAP@50 of 0.92891 for object detection, while YOLOv11n-Seg reaches a segmentation precision of 0.91587 and mAP@50 of 0.79795. For reasoning generation, we tested a range of large vision-language models (LVLMs) in both zero-shot and few-shot settings. Our fine-tuned Mosquito-LLaMA3-8B model achieved the best results, with a final loss of 0.0028, a BLEU score of 54.7, BERTScore of 0.91, and ROUGE-L of 0.85. This dataset and model framework emphasize the theme "Prevention is Better than Cure", showcasing how AI-based detection can proactively address mosquito-borne disease risks. The dataset and implementation code are publicly available at GitHub: https://github.com/adnanul-islam-jisun/VisText-Mosquito

Farshid Rayhan, Sajid Ahmed, Asif Mahbub et al.

Class imbalance problem has been a challenging research problem in the fields of machine learning and data mining as most real life datasets are imbalanced. Several existing machine learning algorithms try to maximize the accuracy classification by correctly identifying majority class samples while ignoring the minority class. However, the concept of the minority class instances usually represents a higher interest than the majority class. Recently, several cost sensitive methods, ensemble models and sampling techniques have been used in literature in order to classify imbalance datasets. In this paper, we propose MEBoost, a new boosting algorithm for imbalanced datasets. MEBoost mixes two different weak learners with boosting to improve the performance on imbalanced datasets. MEBoost is an alternative to the existing techniques such as SMOTEBoost, RUSBoost, Adaboost, etc. The performance of MEBoost has been evaluated on 12 benchmark imbalanced datasets with state of the art ensemble methods like SMOTEBoost, RUSBoost, Easy Ensemble, EUSBoost, DataBoost. Experimental results show significant improvement over the other methods and it can be concluded that MEBoost is an effective and promising algorithm to deal with imbalance datasets. The python version of the code is available here: https://github.com/farshidrayhanuiu/

Proma Hossain Progga, Md. Jobayer Rahman, Swapnil Biswas et al.

Gait recognition is a significant biometric technique for person identification, particularly in scenarios where other physiological biometrics are impractical or ineffective. In this paper, we address the challenges associated with gait recognition and present a novel approach to improve its accuracy and reliability. The proposed method leverages advanced techniques, including sequential gait landmarks obtained through the Mediapipe pose estimation model, Procrustes analysis for alignment, and a Siamese biGRU-dualStack Neural Network architecture for capturing temporal dependencies. Extensive experiments were conducted on large-scale cross-view datasets to demonstrate the effectiveness of the approach, achieving high recognition accuracy compared to other models. The model demonstrated accuracies of 95.7%, 94.44%, 87.71%, and 86.6% on CASIA-B, SZU RGB-D, OU-MVLP, and Gait3D datasets respectively. The results highlight the potential applications of the proposed method in various practical domains, indicating its significant contribution to the field of gait recognition.

Sahid Hossain Mustakim, S M Jishanul Islam, Ummay Maria Muna et al.

Multimodal Large Language Models (MLLMs) are increasingly used for content moderation, yet their robustness in short-form video contexts remains underexplored. Current safety evaluations often rely on unimodal attacks, failing to address combined attack vulnerabilities. In this paper, we introduce a comprehensive framework for evaluating the tri-modal safety of MLLMs. First, we present the Short-Video Multimodal Adversarial (SVMA) dataset, comprising diverse short-form videos with human-guided synthetic adversarial attacks. Second, we propose ChimeraBreak, a novel tri-modal attack strategy that simultaneously challenges visual, auditory, and semantic reasoning pathways. Extensive experiments on state-of-the-art MLLMs reveal significant vulnerabilities with high Attack Success Rates (ASR). Our findings uncover distinct failure modes, showing model biases toward misclassifying benign or policy-violating content. We assess results using LLM-as-a-judge, demonstrating attack reasoning efficacy. Our dataset and findings provide crucial insights for developing more robust and safe MLLMs.

Osman Goni, Himadri Saha Arka, Mithun Halder et al.

Recent advances in Generative Adversarial Networks (GANs) have demonstrated their capability for producing high-quality images. However, a significant challenge remains mode collapse, which occurs when the generator produces a limited number of data patterns that do not reflect the diversity of the training dataset. This study addresses this issue by proposing a number of architectural changes aimed at increasing the diversity and stability of GAN models. We start by improving the loss function with Wasserstein loss and Gradient Penalty to better capture the full range of data variations. We also investigate various network architectures and conclude that ResNet significantly contributes to increased diversity. Building on these findings, we introduce HingeRLC-GAN, a novel approach that combines RLC Regularization and the Hinge loss function. With a FID Score of 18 and a KID Score of 0.001, our approach outperforms existing methods by effectively balancing training stability and increased diversity.

Md. Tarek Hasan, Arifa Akter, Mohammad Nazmush Shamael et al.

Dropout is an effective strategy for the regularization of deep neural networks. Applying tabu to the units that have been dropped in the recent epoch and retaining them for training ensures diversification in dropout. In this paper, we improve the Tabu Dropout mechanism for training deep neural networks in two ways. Firstly, we propose to use tabu tenure, or the number of epochs a particular unit will not be dropped. Different tabu tenures provide diversification to boost the training of deep neural networks based on the search landscape. Secondly, we propose an adaptive tabu algorithm that automatically selects the tabu tenure based on the training performances through epochs. On several standard benchmark datasets, the experimental results show that the adaptive tabu dropout and tabu tenure dropout diversify and perform significantly better compared to the standard dropout and basic tabu dropout mechanisms.

Chowdhury Mofizur Rahman, Mahbub E Sobhani, Anika Tasnim Rodela et al.

Text compression shrinks textual data while keeping crucial information, eradicating constraints on storage, bandwidth, and computational efficacy. The integration of lossless compression techniques with transformer-based text decompression has received negligible attention, despite the increasing volume of English text data in communication. The primary barrier in advancing text compression and restoration involves optimizing transformer-based approaches with efficient pre-processing and integrating lossless compression algorithms, that remained unresolved in the prior attempts. Here, we propose a transformer-based method named RejuvenateForme for text decompression, addressing prior issues by harnessing a new pre-processing technique and a lossless compression method. Our meticulous pre-processing technique incorporating the Lempel-Ziv-Welch algorithm achieves compression ratios of 12.57, 13.38, and 11.42 on the BookCorpus, EN-DE, and EN-FR corpora, thus showing state-of-the-art compression ratios compared to other deep learning and traditional approaches. Furthermore, the RejuvenateForme achieves a BLEU score of 27.31, 25.78, and 50.45 on the EN-DE, EN-FR, and BookCorpus corpora, showcasing its comprehensive efficacy. In contrast, the pre-trained T5-Small exhibits better performance over prior state-of-the-art models.

Umme Abira Azmary, MD Ikramul Kayes, Swakkhar Shatabda et al.

Question-Answering (QA) models for low-resource languages like Bangla face challenges due to limited annotated data and linguistic complexity. A key issue is determining whether models rely more on pre-encoded (parametric) knowledge or contextual input during answer generation, as existing Bangla QA datasets lack the structure required for such analysis. We introduce BanglaCQA, the first Counterfactual QA dataset in Bangla, by extending a Bangla dataset while integrating counterfactual passages and answerability annotations. In addition, we propose fine-tuned pipelines for encoder-decoder language-specific and multilingual baseline models, and prompting-based pipelines for decoder-only LLMs to disentangle parametric and contextual knowledge in both factual and counterfactual scenarios. Furthermore, we apply LLM-based and human evaluation techniques that measure answer quality based on semantic similarity. We also present a detailed analysis of how models perform across different QA settings in low-resource languages, and show that Chain-of-Thought (CoT) prompting reveals a uniquely effective mechanism for extracting parametric knowledge in counterfactual scenarios, particularly in decoder-only LLMs. Our work not only introduces a novel framework for analyzing knowledge sources in Bangla QA but also uncovers critical findings that open up broader directions for counterfactual reasoning in low-resource language settings.

Tanjim Islam Riju, Shuchismita Anwar, Saman Sarker Joy et al.

We propose a two-stage multimodal framework that enhances disease classification and region-aware radiology report generation from chest X-rays, leveraging the MIMIC-Eye dataset. In the first stage, we introduce a gaze-guided contrastive learning architecture for disease classification. It integrates visual features, clinical labels, bounding boxes, and radiologist eye-tracking signals and is equipped with a novel multi-term gaze-attention loss combining MSE, KL divergence, correlation, and center-of-mass alignment. Incorporating fixations improves F1 score from 0.597 to 0.631 (+5.70%) and AUC from 0.821 to 0.849 (+3.41%), while also improving precision and recall, highlighting the effectiveness of gaze-informed attention supervision. In the second stage, we present a modular report generation pipeline that extracts confidence-weighted diagnostic keywords, maps them to anatomical regions using a curated dictionary constructed from domain-specific priors, and generates region-aligned sentences via structured prompts. This pipeline improves report quality as measured by clinical keyword recall and ROUGE overlap. Our results demonstrate that integrating gaze data improves both classification performance and the interpretability of generated medical reports.

Ahmed Rafi Hasan, Niloy Kumar Kundu, Saad Hasan et al.

The Alternate Wetting and Drying (AWD) method is a rice-growing water management technique promoted as a sustainable alternative to Continuous Flooding (CF). Climate change has placed the agricultural sector in a challenging position, particularly as global water resources become increasingly scarce, affecting rice production on irrigated lowlands. Rice, a staple food for over half of the world's population, demands significantly more water than other major crops. In Bangladesh, Boro rice, in particular, requires considerable water inputs during its cultivation. Traditionally, farmers manually measure water levels, a process that is both time-consuming and prone to errors. While ultrasonic sensors offer improvements in water height measurement, they still face limitations, such as susceptibility to weather conditions and environmental factors. To address these issues, we propose a novel approach that automates water height measurement using computer vision, specifically through a convolutional neural network (CNN). Our attention-based architecture achieved an $R^2$ score of 0.9885 and a Mean Squared Error (MSE) of 0.2766, providing a more accurate and efficient solution for managing AWD systems.

Md. Rayhan Ahmed, Salekul Islam, Ph. D et al.

In this paper, we propose an ensemble of deep neural networks along with data augmentation (DA) learned using effective speech-based features to recognize emotions from speech. Our ensemble model is built on three deep neural network-based models. These neural networks are built using the basic local feature acquiring blocks (LFAB) which are consecutive layers of dilated 1D Convolutional Neural networks followed by the max pooling and batch normalization layers. To acquire the long-term dependencies in speech signals further two variants are proposed by adding Gated Recurrent Unit (GRU) and Long Short Term Memory (LSTM) layers respectively. All three network models have consecutive fully connected layers before the final softmax layer for classification. The ensemble model uses a weighted average to provide the final classification. We have utilized five standard benchmark datasets: TESS, EMO-DB, RAVDESS, SAVEE, and CREMA-D for evaluation. We have performed DA by injecting Additive White Gaussian Noise, pitch shifting, and stretching the signal level to generalize the models, and thus increasing the accuracy of the models and reducing the overfitting as well. We handcrafted five categories of features: Mel-frequency cepstral coefficients, Log Mel-Scaled Spectrogram, Zero-Crossing Rate, Chromagram, and statistical Root Mean Square Energy value from each audio sample. These features are used as the input to the LFAB blocks that further extract the hidden local features which are then fed to either fully connected layers or to LSTM or GRU based on the model type to acquire the additional long-term contextual representations. LFAB followed by GRU or LSTM results in better performance compared to the baseline model. The ensemble model achieves the state-of-the-art weighted average accuracy in all the datasets.

Md. Mahinur Rashid, Hasin Kawsar Jahan, Annysha Huzzat et al.

Generation of scientific visualization from analytical natural language text is a challenging task. In this paper, we propose Text2Chart, a multi-staged chart generator method. Text2Chart takes natural language text as input and produce visualization as two-dimensional charts. Text2Chart approaches the problem in three stages. Firstly, it identifies the axis elements of a chart from the given text known as x and y entities. Then it finds a mapping of x-entities with its corresponding y-entities. Next, it generates a chart type suitable for the given text: bar, line or pie. Combination of these three stages is capable of generating visualization from the given analytical text. We have also constructed a dataset for this problem. Experiments show that Text2Chart achieves best performances with BERT based encodings with LSTM models in the first stage to label x and y entities, Random Forest classifier for the mapping stage and fastText embedding with LSTM for the chart type prediction. In our experiments, all the stages show satisfactory results and effectiveness considering formation of charts from analytical text, achieving a commendable overall performance.

Ruhul Amin, Chowdhury Rafeed Rahman, Md. Sadrul Islam Toaha et al.

DNA N6-methylation (6mA) in Adenine nucleotide is a post replication modification and is responsible for many biological functions. Experimental methods for genome wide 6mA site detection is an expensive and manual labour intensive process. Automated and accurate computational methods can help to identify 6mA sites in long genomes saving significant time and money. Our study develops a convolutional neural network based tool i6mA-CNN capable of identifying 6mA sites in the rice genome. Our model coordinates among multiple types of features such as PseAAC inspired customized feature vector, multiple one hot representations and dinucleotide physicochemical properties. It achieves area under the receiver operating characteristic curve of 0.98 with an overall accuracy of 0.94 using 5 fold cross validation on benchmark dataset. Finally, we evaluate our model on two other plant genome 6mA site identification datasets besides rice. Results suggest that our proposed tool is able to generalize its ability of 6mA site identification on plant genomes irrespective of plant species. Web tool for this research can be found at: https://cutt.ly/Co6KuWG. Supplementary data (benchmark dataset, independent test dataset, comparison purpose dataset, trained model, physicochemical property values, attention mechanism details for motif finding) are available at https://cutt.ly/PpDdeDH.

Ruhul Amin, Chowdhury Rafeed Rahman, Md. Habibur Rahman Sifat et al.

Promoter is a short region of DNA which is responsible for initiating transcription of specific genes. Development of computational tools for automatic identification of promoters is in high demand. According to the difference of functions, promoters can be of different types. Promoters may have both intra and inter class variation and similarity in terms of consensus sequences. Accurate classification of various types of sigma promoters still remains a challenge. We present iPromoter-BnCNN for identification and accurate classification of six types of promoters - sigma24, sigma28, sigma32, sigma38, sigma54, sigma70. It is a Convolutional Neural Network (CNN) based classifier which combines local features related to monomer nucleotide sequence, trimer nucleotide sequence, dimer structural properties and trimer structural properties through the use of parallel branching. We conducted experiments on a benchmark dataset and compared with two state-of-the-art tools to show our supremacy on 5-fold cross-validation. Moreover, we tested our classifier on an independent test dataset. Our proposed tool iPromoter-BnCNN web server is freely available at http://103.109.52.8/iPromoter-BnCNN. The runnable source code can be found at https://colab.research.google.com/drive/1yWWh7BXhsm8U4PODgPqlQRy23QGjF2DZ.

Md. Tashfiqul Bari, Tanvir Hassan, Raisa Tabassum et al.

Autism Spectrum Disorder (ASD) is the area where many researches enduring like Magnetic Resonance Imaging (MRI), called diffusion tensor imaging, Early Start Denver Model (ESDM) to provide an easier life for the people diagnosed. After years and years of combined funding sources from public and private funding, these researches show great promises in recent years. In this paper, we have tried to show a way how children with Down Syndrome Autism can learn through game therapy. These game therapies have shown an immense number of improvements among those children to learn alphabets along with developing their motor skills and memory challenges.

Md Mofijul Islam, Amar Debnath, Tahsin Al Sayeed et al.

Deep Learning algorithms are often used as black box type learning and they are too complex to understand. The widespread usability of Deep Learning algorithms to solve various machine learning problems demands deep and transparent understanding of the internal representation as well as decision making. Moreover, the learning models, trained on sequential data, such as audio and video data, have intricate internal reasoning process due to their complex distribution of features. Thus, a visual simulator might be helpful to trace the internal decision making mechanisms in response to adversarial input data, and it would help to debug and design appropriate deep learning models. However, interpreting the internal reasoning of deep learning model is not well studied in the literature. In this work, we have developed a visual interactive web application, namely d-DeVIS, which helps to visualize the internal reasoning of the learning model which is trained on the audio data. The proposed system allows to perceive the behavior as well as to debug the model by interactively generating adversarial audio data point. The web application of d-DeVIS is available at ddevis.herokuapp.com.

Farshid Rayhan, Sajid Ahmed, Zaynab Mousavian et al.

The task of drug-target interaction prediction holds significant importance in pharmacology and therapeutic drug design. In this paper, we present FRnet-DTI, an auto encoder and a convolutional classifier for feature manipulation and drug target interaction prediction. Two convolutional neural neworks are proposed where one model is used for feature manipulation and the other one for classification. Using the first method FRnet-1, we generate 4096 features for each of the instances in each of the datasets and use the second method, FRnet-2, to identify interaction probability employing those features. We have tested our method on four gold standard datasets exhaustively used by other researchers. Experimental results shows that our method significantly improves over the state-of-the-art method on three of the four drug-target interaction gold standard datasets on both area under curve for Receiver Operating Characteristic(auROC) and area under Precision Recall curve(auPR) metric. We also introduce twenty new potential drug-target pairs for interaction based on high prediction scores. Codes Available: https: // github. com/ farshidrayhanuiu/ FRnet-DTI/ Web Implementation: http: // farshidrayhan. pythonanywhere. com/ FRnet-DTI/

Farshid Rayhan, Sajid Ahmed, Asif Mahbub et al.

Class imbalance classification is a challenging research problem in data mining and machine learning, as most of the real-life datasets are often imbalanced in nature. Existing learning algorithms maximise the classification accuracy by correctly classifying the majority class, but misclassify the minority class. However, the minority class instances are representing the concept with greater interest than the majority class instances in real-life applications. Recently, several techniques based on sampling methods (under-sampling of the majority class and over-sampling the minority class), cost-sensitive learning methods, and ensemble learning have been used in the literature for classifying imbalanced datasets. In this paper, we introduce a new clustering-based under-sampling approach with boosting (AdaBoost) algorithm, called CUSBoost, for effective imbalanced classification. The proposed algorithm provides an alternative to RUSBoost (random under-sampling with AdaBoost) and SMOTEBoost (synthetic minority over-sampling with AdaBoost) algorithms. We evaluated the performance of CUSBoost algorithm with the state-of-the-art methods based on ensemble learning like AdaBoost, RUSBoost, SMOTEBoost on 13 imbalance binary and multi-class datasets with various imbalance ratios. The experimental results show that the CUSBoost is a promising and effective approach for dealing with highly imbalanced datasets.

Sajid Ahmed, Farshid Rayhan, Asif Mahbub et al.

The problem of class imbalance along with class-overlapping has become a major issue in the domain of supervised learning. Most supervised learning algorithms assume equal cardinality of the classes under consideration while optimizing the cost function and this assumption does not hold true for imbalanced datasets which results in sub-optimal classification. Therefore, various approaches, such as undersampling, oversampling, cost-sensitive learning and ensemble based methods have been proposed for dealing with imbalanced datasets. However, undersampling suffers from information loss, oversampling suffers from increased runtime and potential overfitting while cost-sensitive methods suffer due to inadequately defined cost assignment schemes. In this paper, we propose a novel boosting based method called LIUBoost. LIUBoost uses under sampling for balancing the datasets in every boosting iteration like RUSBoost while incorporating a cost term for every instance based on their hardness into the weight update formula minimizing the information loss introduced by undersampling. LIUBoost has been extensively evaluated on 18 imbalanced datasets and the results indicate significant improvement over existing best performing method RUSBoost.

Rezaul Karim, Md. Momin Al Aziz, Swakkhar Shatabda et al.

Protein tertiary structure defines its functions, classification and binding sites. Similar structural characteristics between two proteins often lead to the similar characteristics thereof. Determining structural similarity accurately in real time is a crucial research issue. In this paper, we present a novel and effective scoring scheme that is dependent on novel features extracted from protein alpha carbon distance matrices. Our scoring scheme is inspired from pattern recognition and computer vision. Our method is significantly better than the current state of the art methods in terms of family match of pairs of protein structures and other statistical measurements. The effectiveness of our method is tested on standard benchmark structures. A web service is available at http://research.buet.ac.bd:8080/Comograd/score.html where you can get the similarity measurement score between two protein structures based on our method.

Shanjida Khatun, Hasib Ul Alam, Swakkhar Shatabda

Working with exhaustive search on large dataset is infeasible for several reasons. Recently, developed techniques that made pattern set mining feasible by a general solver with long execution time that supports heuristic search and are limited to small datasets only. In this paper, we investigate an approach which aims to find diverse set of patterns using genetic algorithm to mine diverse frequent patterns. We propose a fast heuristic search algorithm that outperforms state-of-the-art methods on a standard set of benchmarks and capable to produce satisfactory results within a short period of time. Our proposed algorithm uses a relative encoding scheme for the patterns and an effective twin removal technique to ensure diversity throughout the search.

Muktadir Hossain, Tajkia Tasnim, Swakkhar Shatabda et al.

Local search methods can quickly find good quality solutions in cases where systematic search methods might take a large amount of time. Moreover, in the context of pattern set mining, exhaustive search methods are not applicable due to the large search space they have to explore. In this paper, we propose the application of stochastic local search to solve the pattern set mining. Specifically, to the task of concept learning. We applied a number of local search algorithms on a standard benchmark instances for pattern set mining and the results show the potentials for further exploration.

Mohammad Murtaza Mahmud, Swakkhar Shatabda, Mohammad Nurul Huda

Automated Theorem Proving (ATP) is an established branch of Artificial Intelligence. The purpose of ATP is to design a system which can automatically figure out an algorithm either to prove or disprove a mathematical claim, on the basis of a set of given premises, using a set of fundamental postulates and following the method of logical inference. In this paper, we propose GraATP, a generalized framework for automated theorem proving in plane geometry. Our proposed method translates the geometric entities into nodes of a graph and the relations between them as edges of that graph. The automated system searches for different ways to reach the conclusion for a claim via graph traversal by which the validity of the geometric theorem is examined.

Md. Lisul Islam, Novia Nurain, Swakkhar Shatabda et al.

Graph partitioning, a well studied problem of parallel computing has many applications in diversified fields such as distributed computing, social network analysis, data mining and many other domains. In this paper, we introduce FGPGA, an efficient genetic approach for producing feasible graph partitions. Our method takes into account the heterogeneity and capacity constraints of the partitions to ensure balanced partitioning. Such approach has various applications in mobile cloud computing that include feasible deployment of software applications on the more resourceful infrastructure in the cloud instead of mobile hand set. Our proposed approach is light weight and hence suitable for use in cloud architecture. We ensure feasibility of the partitions generated by not allowing over-sized partitions to be generated during the initialization and search. Our proposed method tested on standard benchmark datasets significantly outperforms the state-of-the-art methods in terms of quality of partitions and feasibility of the solutions.

Md. Lisul Islam, Swakkhar Shatabda, M. Sohel Rahman

Nuclear Magnetic Resonance (NMR) Spectroscopy is a widely used technique to predict the native structure of proteins. However, NMR machines are only able to report approximate and partial distances between pair of atoms. To build the protein structure one has to solve the Euclidean distance geometry problem given the incomplete interval distance data produced by NMR machines. In this paper, we propose a new genetic algorithm for solving the Euclidean distance geometry problem for protein structure prediction given sparse NMR data. Our genetic algorithm uses a greedy mutation operator to intensify the search, a twin removal technique for diversification in the population and a random restart method to recover stagnation. On a standard set of benchmark dataset, our algorithm significantly outperforms standard genetic algorithms.

Rezaul Karim, Mohd. Momin Al Aziz, Swakkhar Shatabda et al.

Due to the advancements in technology number of entries in the structural database of proteins are increasing day by day. Methods for retrieving protein tertiary structures from this large database is the key to comparative analysis of structures which plays an important role to understand proteins and their function. In this paper, we present fast and accurate methods for the retrieval of proteins from a large database with tertiary structures similar to a query protein. Our proposed methods borrow ideas from the field of computer vision. The speed and accuracy of our methods comes from the two newly introduced features, the co-occurrence matrix of the oriented gradient and pyramid histogram of oriented gradient and from the use of Euclidean distance as the distance measure. Experimental results clearly indicate the superiority of our approach in both running time and accuracy. Our method is readily available for use from this website: http://research.buet.ac.bd:8080/Comograd/.

Swakkhar Shatabda, M. A. Hakim Newton, Duc Nghia Pham et al.

Protein structure prediction based on Hydrophobic-Polar energy model essentially becomes searching for a conformation having a compact hydrophobic core at the center. The hydrophobic core minimizes the interaction energy between the amino acids of the given protein. Local search algorithms can quickly find very good conformations by moving repeatedly from the current solution to its "best" neighbor. However, once such a compact hydrophobic core is found, the search stagnates and spends enormous effort in quest of an alternative core. In this paper, we attempt to restructure segments of a conformation with such compact core. We select one large segment or a number of small segments and apply exhaustive local search. We also apply a mix of heuristics so that one heuristic can help escape local minima of another. We evaluated our algorithm by using Face Centered Cubic (FCC) Lattice on a set of standard benchmark proteins and obtain significantly better results than that of the state-of-the-art methods.