Paper Submission - NPS Australia Submission System

Driving Sustainable Social Transformation in Bangladesh: Social Business Practice in Rural Communities: A Mixed-Methods Study of Five Grameen-Affiliated Social Business Ventures

The paper contributes theoretically by articulating a two-phased model linking operational drivers to sustainable, self-reinforcing systemic outcomes, and empirically by providing cross-case evidence from Grameen-affiliated SBVs in rural Bangladesh. Read alongside Quilloy, Newman, and Pyman’s (2023) systematic call for research connecting organisational-level antecedents to social-enterprise impact outcomes, the present study’s Phase I-to-Phase II linkage offers one concrete, empirically grounded answer to that call within a single, institutionally coherent ecosystem. Practically, findings indicate that SBVs are most likely to sustain their impact over time when they combine product-market fit, local capacity building, adaptive management, and stakeholder partnerships — the combination most closely associated, in this study, with durable rather than transient outcomes.

Association Rule Mining in Picture Fuzzy Data Mining

In the light of picture fuzzy set, the Apriori algorithm and Frequent Pattern growth algorithm have been used to generate the picture fuzzy rules and FP Tree and applied in market basket analysis with illustrations.

Evaluating Socially Responsible Human-AI Interaction: An Investigation of AI Accommodation to Diverse Human Communication Styles

The recent advancements in conversational AI have transformed the mode of human-AI interaction into a more personalized one. The impact of accommodation of various styles of communication on the part of AI on socially responsible human-AI communication has not been investigated yet sufficiently enough. This paper explores the question of how conversational AI accommodates the linguistic, tonal, and emotional style of users. Moreover, the possibility of accommodating the communication style of AI in order to ensure socially responsible human-AI communication is explored as well. Quantitative data were collected via a structured questionnaire completed by 41 graduate and undergraduate students. According to the results, the respondents have rather positive attitudes regarding AI adaptation of vocabulary use, level of formality, tonal style, and emotional response to the user’s communication style. At the same time, the respondents noted the restrictions of style accommodation and pointed out that there are situations in which style matching is inappropriate and could harm socially responsible human-AI communication. From the obtained results, it can be concluded that adaptive AI will facilitate the process of making communication more accessible and inclusive for users of various linguistic levels.

View-Aware Reliability Framework for Pneumonia Diagnosis with Calibration Portability, Selective Risk, and XAI Auditing

Deep learning models for pneumonia detection from chest radiographs are usually assessed with discrimination metrics, yet comparable discrimination does not guarantee that predicted probabilities remain trustworthy across acquisition views or institutions. This work evaluates a DenseNet121 pneumonia classifier under anteroposterior (AP) and posteroanterior (PA) view variation and cross-dataset shift. A patient-level cohort from NIH ChestX-ray14 was used for development and internal testing, and an independent CheXpert cohort was used for external testing with no retraining, calibration refitting, or threshold adjustment. Global temperature scaling was compared with AP/PA-specific temperature scaling, predictive entropy supported uncertainty-guided selective prediction, and Grad-CAM provided a qualitative reliability audit. The classifier reached AUROCs of 0.7958 on NIH and 0.7859 on CheXpert, indicating relatively stable ranking behavior. Reliability was more sensitive to the shift. View-aware calibration lowered internal expected calibration error from 0.0460 to 0.0361, but the source-fitted temperatures did not improve on raw CheXpert probabilities (0.0822 versus 0.0813). The entropy threshold preserved coverage externally (90.10% to 88.67%) while the area under the riskcoverage curve rose from 0.1285 to 0.1564. PA radiographs showed lower discrimination and poorer calibration than AP radiographs in both cohorts. These results show that calibration and selective reliability can degrade while AUROC appears stable, supporting view- and domain-aware reliability evaluation before cross-dataset deployment.

Environmental Sustainability and Energy Awareness among University Students in Bangladesh: An Empirical Study

Renewable energy is developing gradually in Bangladesh to address climate change, environmental degradation and expanding energy demand. However, the success of this progress needs both technological support and public awareness and adaptation of sustainable energy progress. This study investigates renewable energy and environmental awareness among university students in Bangladesh to examine their behavioral intention towards energy adaptation and sustainable energy trials. The study employed a quantitative cross-sectional method and data were collected using questionnaire surveys in a selected university. The findings showed that there is a clear distinction between overall environmental understanding and detailed energy literacy. Participants have a strong concern about environmental awareness such as in climate change, pollution but they lack understanding technology with the use of energy production, solar systems and renewable energy policies. Moreover, the findings also indicated a positive concern among students toward energy conservation, sustainable lifestyles, renewable-energy adoption, and Solar Home Systems (SHS). The study argues that university students hold a promising backbone to support Bangladesh’s energy transition and to achieve this, their general knowledge must be transformed into informal and practical action. To obliterate the gap in general knowledge the study recommends strengthening education, institutional engagement, and policy commitment to remodel students’ awareness and intention into sustainable energy practices.

Sustainable Consumption and E-Waste Reduction: An Analysis of the Jargon Barrier in Consumer Electronics Readability for Right-to-Repair Practices

Every consumer electronics replacement cycle is driven by real human choices, yet shaped entirely by official product instructions. However, previous studies on environmental sustainability have predominantly focused on industrial hardware manufacturing and post-consumer recycling while largely overlooking how corporate linguistic practices undermine consumer repair agency. This study aims to investigate the jargon barriers and readability measures in consumer electronics manuals that hinder right-to-repair initiatives. To conduct this study, a mixed-methods approach was employed, integrating quantitative surveys of everyday consumers, qualitative interviews with consumers and repair technicians, and readability measures. The findings reflect that consumers face difficulties in carrying out regular maintenance due to complicated technical language, lack of localized Bangla translations, and insufficient visual instructions in product manuals. Consequently, users frequently replace their devices within short lifespans. However, consumers conveyed that clear and understandable instructions build their confidence to attempt basic self-repair. Framing through Communication Accommodation Theory (CAT), the study proposes the application of plain and localized language in national policies and corporate manuals to increase device lifespans. The authors of this paper encourage more studies to explore how simplified and clear technical manuals can help consumers repair devices and support sustainability in tech markets through the reduction of e-waste.

Developing a Financial Innovation Framework for Sustainable Banking Performance: A Conceptual Model for Bangladesh

The adoption of financial innovation should be viewed by banks in Bangladesh as a strategic process that extends beyond technological implementation to encompass organizational transformation and sustainable performance enhancement. The proposed conceptual framework can serve as a valuable guide for bank managers, policymakers, regulators, and financial institutions in understanding the role of financial innovations in improving sustainable banking performance. By identifying the key dimensions of financial innovation i.e. Internet Banking, ATM Services, and Credit
Card Service, the framework enables stakeholders to evaluate innovation priorities, allocate resources more effectively, and formulate strategies that enhance profitability, operational efficiency, and long-term competitiveness. Furthermore, the model provides a context-specific foundation for assessing innovation-driven sustainable performance within the banking sector of Bangladesh and other developing economies experiencing similar technological and institutional challenges.

Optimal Ramp-Rate Control Using Battery Energy Storage Systems (BESS) for Utility-Scale Solar Power Plants in Bangladesh

This research focuses on the prospectively of integrating Battery Energy Storage Systems (BESS) with existing and new solar power plants to level rapid fluctuations in PV generation and enhance grid reliability. It also pinpoints technical and regulatory measures to speed up the deployment of BESS, outlines the challenges the national grid will face if large numbers of BESS are deployed, and offers a general perspective on the financial hurdles involved. The results should help the better integration of renewable energies and inform planning and investment in energy storage technologies

Design and Implementation of a Low-Cost Portable Oscilloscope Using Arduino Nano

This work contributes a validated design and implementation of a low-cost portable oscilloscope using Arduino Nano and an OLED display, offering an accessible alternative to expensive laboratory instruments. By integrating signal conditioning, ADC conversion, embedded processing, and real-time visualization, the prototype demonstrates how microcontroller-based systems can effectively support educational experiments, basic circuit debugging, and sensor signal observation. The research highlights the feasibility of achieving waveform acquisition and display at low frequencies (~1 kHz) in a compact and inexpensive package, thereby bridging the gap between theoretical learning and practical measurement tools. It also establishes a foundation for future improvements such as faster sampling, improved triggering, and waveform storage, making it a stepping stone for further work in low-cost instrumentation and embedded measurement systems.

Meta-Learning for DeepFake Detection: An Ensemble-Based Approach

This study proposes an ensemble-based
deepfake detection framework that combines multiple pretrained
convolutional neural networks (CNNs) using a trainable meta-
classifier. Experiments were conducted using the SDFVD2.0 [1]
dataset containing 927 videos (456 real and 471 fake), from
which approximately 2,400 facial images were extracted and
balanced across classes. Three CNN architectures—ResNet101,
DenseNet201, and GoogLeNet—were fine-tuned for binary clas-
sification. Instead of relying on a single model, their outputs were
fused through a meta-learning classifier to improve robustness
and generalization. The proposed ensemble approach achieved
an accuracy of 98.10% on a validation set of 480 images, with
strong precision (97.50%), recall (98.73%), F1-score (98.11%),
and an AUC of 0.992.

An Edge-Cloud Smart Classroom System for Energy Efficiency with Intelligent Classroom and Exam Modes

This paper presented a classroom monitoring and automa-
tion system developed using ESP32, Firebase and computer

vision. The prototype controlled temperature, smart light, door

actuation, emergency alerts, student counting and exam mon-
itoring on a single platform. Testing proved that the system

responded to sensor events and updated Firebase without
noticeable delay. The student counting achieved 92% accuracy
and cheating detection achieved 85% accuracy during exam
mode trials.
The result demonstrates the potential of useful classroom
automation using low-cost hardware and common software
tools. For future work, we plan to improve the placement of

cameras for improved detection accuracy, and to run light-
weight vision models directly on the embedded hardware, thus

removing the dependency on a processing machine external
to the hardware. In the future we will optimise camera
placements for better detection accuracy and run lightweight
vision models directly on the embedded hardware, without an
external processing machine.

Hybrid Reconstruction Approaches for Occluded Regions in Multi-Person Gait Recognition

This work introduces a hybrid two-stage framework for reconstructing occluded body landmarks in multi-person gait recognition, addressing a major barrier to deploying gait biometrics in real-world surveillance. Unlike prior methods that use either numerical interpolation or optimization alone, the proposed approach fuses cubic spline interpolation with optimization-based refinement (ANN, ALS, and PSO-NN), where interpolation establishes a biomechanically consistent baseline that optimization then refines. Evaluated on the benchmark SMVDU dataset and a newly collected RUET-MG dataset across controlled occlusion levels of 10–40%, the PSO-NN variant consistently achieves the lowest reconstruction error (X-MSE of 0.103 on SMVDU-MG). This represents an improvement of over 97% against optimization-only methods and over 75% against interpolation-only methods, establishing the fused pipeline as a practical, occlusion-robust solution for outdoor multi-person gait analysis.

A Benchmark of Fake Review Detection Using Machine Learning in E-Commerce

The increasing use of online reviews has made reliable review information essential for consumer decision-making, while deceptive reviews reduce trust in e-commerce platforms. This study addresses fake review detection in Banglish, a code-mixed form of Bangla and English for which benchmark datasets and specialized detection frameworks are limited. A dataset of 25,588 reviews was developed and annotated into fake and not_fake classes, capturing Bangla script, Romanized Bangla, English, transliteration, and informal code-mixing. A hybrid BanglaBERT-BiLSTM model is proposed, where BanglaBERT learns contextual semantic representations and BiLSTM captures sequential dependencies. The proposed system achieved 96.82% accuracy, 96.83% precision, 96.82% recall, 96.44% F1-score, and 99.47% ROC-AUC on the unseen test set. It outperformed the evaluated traditional and transformer-based baselines. LIME and SHAP analyses further provide interpretable evidence about the textual features influencing predictions. The dataset and benchmark provide a foundation for future research in Banglish NLP and deceptive review detection.

Regression-Based Prediction of Under-Five Child Malnutrition in Bangladesh Using Machine Learning on BDHS 2022

—Child malnutrition remains a major public-health
challenge in Bangladesh and is associated with increased mor
bidity, impaired physical growth, delayed cognitive development,
and elevated mortality risk among children under five years of
age. Conventional nutritional assessment relies on anthropomet
ric measurements and threshold-based interpretation, but these
approaches may not fully capture the combined influence of
child-level, maternal, household, socioeconomic, and geographic
factors. This paper presents a regression-based machine-learning
framework for predicting the weight-for-height Z-score (WHZ)
of Bangladeshi children aged 0–59 months using the Bangladesh
Demographic and Health Survey 2022 children recode dataset.
After applying inclusion and quality-control criteria, the final
analytic dataset contained 4,105 records and 24 selected at
tributes. The study compares Linear Regression, Decision Tree
Regressor, Random Forest Regressor, and K-Nearest Neighbors
Regressor using mean squared error, root mean squared error,
mean absolute error, and coefficient of determination. Feature
engineering was applied to derive child body mass index, weight
to-age ratio, height-to-age ratio, and age-group variables. In
addition, K-means clustering with K = 3 was used to explore
nutritional-risk grouping in an unsupervised manner. The Ran
dom Forest Regressor achieved the best validation performance
with R2 = 0.9339, RMSE = 0.3166, and MAE = 0.2128, and
maintained strong test performance with R2 = 0.9249, RMSE
=0.3558, and MAE = 0.2374. The results suggest that ensemble
regression combined with anthropometric feature engineering
can provide an effective data-driven approach for estimating
WHZ scores from survey variables, although causal inference
and deployment-level conclusions require further validation.
Index Terms—child malnutrition, WHZ score, machine learn
ing, Random Forest, regression, Bangladesh DHS, anthropomet
ric feature engineering, K-means clustering

AI-Based Greenwashing Detection on Digital Platforms Using Multi-Modal Deep Learning

With the growing trend of sustainability comes greenwashing, where companies make green claims to attract green-minded consumers. Environmental claims misuse is on the rise, resulting in greenwashing, where companies misrepresent their products or activities as good for the planet. It is crucial to detect greenwashing to maintain customer trust, particularly in this digital age, where confusion and misleading information are prevalent. This work presents DL approaches to detect greenwashing with a focus on scrutinizing customer skepticism. We recommend the use of DL in the identification of environmental deception claims and can further trace consumer skepticism in digital sources (social media, e-commerce sites, business communications). CNN is built using convolution for hierarchical feature extraction and capturing spatial relations within the input data. GoogleNet refines this with inception modules for multi-scale feature extraction, together with global average pooling to promote feature economy. VGG19, known for its deep structure, is often used in complicated pattern recognition. These models learn from an ESG dataset crafted to embody consumer skepticism, and thus can recognize and discern genuine environmental claims from greenwashing. We have implemented the DL models that outperform the baseline models on the ESG dataset with accuracy (99.67%, 99.56%, 99.24%), Precision (99.47%, 1.00, 98.43%), Recall (99.73%, 98.94%, 99.73%), and F1 -score (99.60%, 99.47%, 99.08%) for GoogleNet, VGG19-Net, and CNN, respectively. The results demonstrate the effectiveness of DL for understanding consumer sentiment against fraudulent marketing tactics on digital platforms, enabling the detection of greenwashing.

Extending the Horizon: Uncertainty-Aware 4D Scene Generation for Robust Autonomous Driving

4D LiDAR scene generation for autonomous vehicles
is limited by a uniformity fallacy: existing models allocate equal
computational effort across all spatial regions, starving rare and
dynamically critical areas of representational budget. This produces three compounding failures—generation quality degrades
beyond roughly ten frames (approximately 0.5 seconds), outof-distribution detection is performed per-frame and therefore
misses temporal anomalies, and segment-first pipelines lose novel
or occluded objects at the initial detection step. We propose a
hard-first, uncertainty-guided generation framework that schedules capacity to high-entropy regions—dynamic outliers, occlusions, and unseen object classes—before resolving background
geometry. Shannon entropy serves as the routing signal, directing the denoising process to front-load difficult regions while
the model retains maximum representational freedom. Openvocabulary semantic priors are coupled with the uncertainty
routing to label and render novel object classes that segment-first
pipelines discard. We evaluate on nuScenes and report Expected
Calibration Error under out-of-distribution scenarios alongside
temporal-coherence and horizon-length metrics. By prioritizing
the regions where the model is least certain, the framework
extends reliable generation horizons, reduces calibration error
on rare-event scenarios, and enables safer downstream planning.

Impact of Unemployment, Inflation, Renewable Energy Consumption, and Personal Remittances on Sustainable Economic Growth in Bangladesh

The study contributes to the literature by offering a novel, integrated empirical analysis of renewable energy consumption, inflation, unemployment, and personal remittances on GDP growth in Bangladesh—a combination not previously examined together in recent years. It explicitly anchors its investigation within the frameworks of SDG 7 (clean energy) and SDG 8 (decent work and economic growth), thereby linking macroeconomic and green energy variables to sustainable development priorities. Furthermore, by distinguishing between short-run and long-run dynamic effects, the research provides targeted, country-specific policy insights to help Bangladesh.

Property Engineering of Cobalt-Doped Bismuth Ferrite Nanoparticles for Photovoltaic Energy Harvesting

This work provides an integrated, single-study assessment of how 10% cobalt substitution at the Fe (B) site engineers the structural, morphological, optical, and electrical properties of BiFeO₃ nanoparticles — properties that are typically reported in isolation across separate studies. Prepared under identical sol–gel and annealing conditions to enable direct comparison, pristine BFO and BFCO-10 are shown to retain the rhombohedral R3c perovskite framework (with only a trace Bi₂₅FeO₄₀ phase) while Co doping simultaneously contracts the unit cell, raises microstrain, and refines the crystallite size from 65.7 to 49 nm and the particle size from 170 to 128 nm (SEM). Most significantly, Co substitution narrows the direct optical band gap from 2.06 to 1.80 eV — extending visible-light absorption toward the photovoltaic-optimal range — while concurrently lowering AC resistivity and enhancing conductivity through defect-mediated (Fe²⁺/Fe³⁺, Co²⁺/Co³⁺) charge transport. The study’s contribution lies in linking these effects within one coherent structure–morphology–property framework, demonstrating that modest B-site Co doping is an effective single-dopant route to tailor BiFeO₃ for visible-light photovoltaic energy harvesting without disrupting the parent perovskite structure.

Explainable Deep Learning for Credit Risk Assessment: A Data-Driven Decision Support Framework for Financial Institutions

Credit risk assessment is a critical job for financial institutions, as it can affect credit loss, restrict profitable lending opportunities, and raise issues for the institutions related to their operations and/or reputations. Though traditional statistical credit-scoring models are easily understood, they might not be capable of fully revealing complex non-linear interactions among borrower data. It is therefore desirable to explore a different approach to learning the nonlinear interactions in multidimensional credit data, and deep learning is an alternative that has been suggested; it remains a challenge, however, as it is non-textual, which gives rise to issues of managerial acceptability and responsibility, transparency, and justice. This study has incorporated three components of explainable deep learning decision support systems: A multilayer neural network, probability calibration, and Shapley Additive exPlanations (SHAP), along with risk-based decision routing and fairness monitoring. The proof-of-concept study was based on a synthetic population of 30,000 credit applicants who have 19 financial and behavioral variables. The proposed deep neural network achieved an accuracy of 0.9891, an ROC-AUC value of 0.9965, a balanced accuracy of 0.9886, a recall score of 0.9875, and an F1 score of 0.9784. Platt calibration improved its Brier score from 0.0106 to 0.0068 without decreasing discriminating ability. The most significant worldwide risk variables were found by the SHAP study to be recent delinquencies, missing payments, debt-to-income ratio, credit utilization, and length of credit history. Examples presented throughout this study demonstrate that all of these can be integrated into a single decision support system, which includes: predictive modelling, explanation, calibration, human oversight, and fairness audits. The proposed approach does not aim to replace credit officers but rather to give financial institutions clear and transparent evidence of risk that supports a consistent and operationally able lending process.

Cross-Modal Steganography Detection Using Multimodal Large Language Models: A Comparative Study of Zero-Shot Prompting, Classical Steganalysis, and Parameter-Efficient Fine-Tuning

Classical steganalysis tools such as the Spatial Rich Model (SRM) assume stationary natural-image sensor noise, an assumption violated by synthetic imagery from GANs and diffusion models. We investigate whether a multimodal large language model (MLLM), LLaVA-1.6-Mistral-7B, can detect least significant bit (LSB) steganography by reasoning about texture and noise rather than a fixed statistical model. On a 16,000-record dataset built from CIFAR-10 and CelebA covers with LSB embedding at 10%-30% capacity, plus JPEG and Gaussian adversarial variants, we compare zero-shot chain-of-thought prompting, an SRM baseline, and LoRA fine-tuned LLaVA with answer-only label masking on a stratified 200-image test set. Zero shot LLaVA is near-chance (AUC = 0.526). SRM reaches AUC = 0.730; however, its F1 collapses from 0.850 to 0.546 at 5% FPR. LoRA fine-tuning reaches AUC = 0.807 and F1@5%FPR = 0.743, meeting both targets. We show that the decisive factor is loss masking: unmasked training loss is unstable (=4.8), while masking non-answer tokens to −100 yields convergence from 0.117 to 0.022 over 10 epochs. The pipeline runs as a self-correcting SLURM job on four A100 GPUs.

Knowledge Augmented Ambiguity Aware Learning for Multi-Label Emotion Detection

We propose KAME, a Knowledge-Augmented Ambiguity-Aware Learning framework for robust multi-label emotion detection. KAME enriches transformer-based representations with zero-shot semantic priors obtained from an external natural language inference model and introduces entropy-guided ambiguity weighting to reduce the influence of highly uncertain training instances. The framework further incorporates R-Drop consistency regularization, exponential moving averaging, and multi-seed/multi-backbone probability ensembling with development-set threshold optimization. Experiments on English and Chinese multi-label emotion datasets demonstrate strong and balanced performance. KAME achieves a Macro-F1 of 0.650 and Hamming loss of 0.117 on English, while obtaining the best Micro-F1 of 0.673, Macro-F1 of 0.616, and Hamming loss of 0.136 on Chinese among the compared methods. Ablation results further confirm the importance of semantic priors, consistency regularization, EMA, and ensemble aggregation. Overall, KAME provides an effective framework for integrating external semantic knowledge, uncertainty-aware learning, and robust prediction for multi-label emotion recognition.

Design and Implementation of KitchenGuard: A Multi-Sensor IoT Framework for Robust Kitchen Hazard Mitigation with Edge-Cloud Actuation

KitchenGuard is an inexpensive multi-sensor IoT solution for combustible-gas, smoke, thermal and flame detection in the home kitchen. The ESP32 edge controller conducts local sensing, perform hazard fusion and actuate without relying on cloud connectivity and cloud dashboard offers remote visibility. The prototype involves the integration of MQ-series gas/air-quality sensors, temperature-humidity sensing, infrared flame detection, solenoid valve, exhaust fan, water pump, buzzer, and, local display. A normalized Kitchen Hazard Index includes sensor characteristics, flame and extreme gas are hard over-ride conditions. Experimental testing revealed that the temperature dropped from 35.0°C to 32.1°C within 12.2 min after the fan was turned on. The spatial gradient was clearly evident in the gas placement readings: 4095 at the source, 3728 below the cooker, 3191 above the cooker, and ~1690 at both corners three feet away. Project records show hazard response time of 1.65-s, gas isolation time of 1.9-s, successful telemetry over 72 h, and cost of less than US$65, but these are not event by event claims and need to be independently verified. The results show that edge based multi-sensor detection, edge based local mitigation and cloud-based observability are feasible as proof-of-concept. In order to validate, sensor calibration, repeated trials, fault testing and standards-based safety assessment must be performed prior to deployment.

Geospatial Suitability Assessment and Techno-Economic Optimization for Green Hydrogen Production in Coastal Bangladesh

This paper presents a two-stage geospatial and
techno-economic framework for identifying and sizing a renewable
hydrogen production system in Bangladesh. In the
first stage, a Geographic Information System (GIS)-based multicriteria
decision analysis was carried out for Chittagong Division
using seven spatial criteria, namely solar irradiance, wind speed,
slope, land use and land cover, and proximity to water bodies,
roads, and powerlines. The criteria were combined through a
weighted linear combination method under five distinct weighting
scenarios to test the sensitivity of the site ranking to the assumed
priorities. The results show that pairing geospatial suitability screening with technoeconomic system optimization gives a practical, data-driven route
for planning renewable hydrogen projects along the coastal belt
of Bangladesh.

When Three Policies Move at Once: Module Market Evidence for Rooftop Solar Procurement in Bangladesh

The paper contributes to the literature by providing a empirical and policy-linked evaluation of rooftop solar module procurement in Bangladesh following the simultaneous implementation of the Net Metering Guidelines 2025, the June 2026 tariff adjustment, and S.R.O. No. 159 (Customs). First, it establishes a technical baseline by standardizing and filtering a market dataset of 121 photovoltaic module models across 27 brands, characterizing key metrics such as median power density (220.6 W/m²) and temperature-induced efficiency loss at 45°C (6.0%). Second, it quantifies the prosumer incentive compression caused by the June 2026 tariff order, demonstrating that a sharper percentage increase in bulk rates (19.85%) relative to retail rates (16.68%) reduced the self-consumption premium from 44.6% to 40.8%. Third, it reveals a novel “value inversion” phenomenon under net-metering settlement rules, where low-tariff consumer groups (such as lifeline and irrigation users) obtain 29.5% and 20.0% higher financial returns from exporting power than from self-consumption. Finally, by mapping effective import duty mechanics, the study corrects the common assumption of full zero-rating under S.R.O. 159, demonstrating that unexempted VAT leaves tax incidence at 17.00% (a 10.02% reduction), which was ultimately eclipsed by a 34.1% rise in upstream module costs and a net 21.9% increase in landed cost.

Reconciling Published Wave Energy Estimates for Coastal Bangladesh: A Comparative Analysis with MATLAB-Based Resource and Cost Reconstruction

The primary contribution of this paper lies in analyzing literature inconsistencies within the Bay of Bengal wave energy domain as a central focus to establish a unified benchmark. First, it presents a structured comparison of nine existing studies across their methodologies, evaluated periods, reported metrics, conversion technologies, and associated costs. Second, it utilizes a custom MATLAB framework to recompute published sea states under a single standardized definition, cleanly decoupling definitional variance from physical wave variance. Third, the study delivers a targeted, device-level feasibility assessment for the region’s best-documented site, completing a key research gap identified in prior work. Finally, it provides an economic evaluation by scaling internationally benchmarked reference-model costs directly to Bangladesh’s wave climate, explicitly quantifying the cost variance relative to regional estimates.

Numerical Assessment of Tidal Energy Resource and Generation Potential for Coastal Bangladesh

The paper contributes to the literature by reconciling the discrepancies among existing national tidal power surveys in Bangladesh through a systematic reconstruction of their methodology and underlying inputs. First, it reconstructs the four published numerical surveys to reveal implicit parameters such as efficiency, basin area, averaging window, and rotor size. Second, it introduces a unified normalized equation (P = kAR^2) for barrage-type estimates, which collapses a factor-of-eleven disagreement across sites down to just 1.3% by isolating three explicit methodological choices. Third, it restores missing unit data in a key monthly dataset, clarifying the resource’s equinoctial seasonal pattern. Finally, it provides a common-basis comparison between tidal range and tidal stream approaches at shared geographic sites, quantifying the comparative output as an equivalent rotor count.

Experimental Assessment of an IoT Home Energy Management System Integrating Dual Axis Solar Tracking and Net Energy Analysis

The main contribution of this research is the design, implementation, and empirical evaluation of a low-cost (approximately BDT 7,700) IoT-based home energy management prototype that integrates synchronized physical and browser-based control, multi-load current sensing, battery storage, and a two-axis photovoltaic tracker. Beyond proving the system’s operational viability at an accessible price point, the study quantifies the performance gains of the two-axis tracker—which yielded a 35.35% mean-power gain and 2.402 Wh in gross energy over a fixed panel—and critically transforms the commonly unverified assumption of negligible tracking power consumption into a precise experimental threshold: the tracker must consume less than 0.600 W on average to maintain a net-positive energy benefit.

An Intelligent Wearable Framework for Real-Time Harassment Detection Using Machine Learning

The work presents an intelligent wearable harassment-detection framework that integrates FSR-based physical interaction sensing, NRF51822 BLE-enabled real-time communication, and cloud-based machine learning. Its main contribution is a stacking ensemble model combining XGBoost, LightGBM, CatBoost, and MLP, with Logistic Regression as the meta-learner, to improve harassment classification performance. The framework also incorporates SHAP-based explainability to interpret sensor contributions and uses multi-metric performance evaluation to assess the reliability and practical suitability of the proposed detection system.

SINDUK: EMPOWERING SENIOR CITIZENS THROUGH MOBILE TECHNOLOGY IN A RAPIDLY DIGITIZING WORLD

This study’s primary contribution is the design and empirical evaluation of SINDUK, an integrated mobile application for Bangladeshi senior citizens. Unlike existing single-purpose elder-care apps, SINDUK bridges the digital divide by uniquely combining centralized medical records, one-touch emergency security, social connectivity, and personal-legacy archiving into a single, family-inclusive platform grounded in the UTAUT2 model and gerontechnology theory.

Explainable Ensemble Learning for Sales Forecasting and Product Ranking in Retail

The current study explores a framework for retail
sales prediction and product selection based on approximately
50,000 sales transactions. Time-based variables such as lagged
sales values, rolling averages, and cyclic month encoding have
been designed to detect sequential purchases. XGBoost, LightGBM, CatBoost, and a weighted ensemble of the previous algorithms have been tested by means of 5-fold time-sensitive crossvalidation with RMSE as a metric. The lowest RMSE of 3958.76
was achieved by CatBoost, outperforming other models such as
XGBoost (4951.30), the weighted hybrid ensemble (5454.23), and
LightGBM (7986.99). Analysis of SHAP values demonstrated
that the most relevant variable is lag 1. Iterative multi-step
forecasting with horizons of 6 and 12 months showed that ”TShirt” and ”Headphones” are top products, while electronics
ranked among the best. These findings indicate that appropriate
engineering of time-based features, gradient boosting algorithms,
and explainable AI techniques can reliably help to make decisions
on inventory management and marketing, but the current model
is limited due to the lack of external variables and accumulation
of errors during iterative forecast

Who did the Thinking? A cognitive-control conceptual framework for student-GenAI interaction in assessed ICT tasks

This paper refines the analytical focus on the real-time distribution of cognitive control and addresses this diagnostic gap. It conceptualises student-GenAI interaction as dynamic distribution of control across four
distinct dimensions (interpretation, comprehension, thinking, and decision-making), and maps them into an engagement continuum. The framework translates these dimensions into behavioural anchors, proposes triangulation across multi-source to reduce self-report bias, and outlines a validation agenda.

Scenario Based Synthetic Road Accident Risk Prediction using Explainable Ensemble Learning

Road accidents have no single factor to take place.
There are many different variables that affect road accidents
including weather, road, traffic, enforcement, drivers, and many
others. The interaction between these variables influences the
number of road accidents. In addition, there are few crash
database available for Bangladesh, and that’s why we created a
synthetic dataset to conduct our analysis. Our data set includes
9,000 records with 28 different columns, each of which represents
the scenario of the district under consideration in the given
month with particular road, weather, and traffic conditions. Five
columns were dropped since they did not exist before the crash
event happened, giving 22 predictors in total. Random Forest,
Extra Trees, Gradient Boosting, and soft voting were performed
in addition to the previous techniques. The base technique was
Logistic Regression. The maximum accuracy score of 83.00% was
attained by Gradient Boosting. Soft voting gave 82.83%. Linear
baseline had 74.44% accuracy score. The macro-average precision,
recall, and F1 score of the best performing model were 0.8321,
0.8302, and 0.8310 respectively. ROC-AUC scores of High risk,
Low, and Medium risk classes were 0.9785, 0.9686, and 0.8943
respectively. The explainable AI analysis was done, and target
leakage was checked. Three results were observed during audit:
no predictor is a fixed function of another predictor; no threshold
value exists for any class; deleting the composite scenario risk
index reduces the accuracy score by 6.66 points. Conclusion
includes a forecast of the changes in risk factors between 2027
and 2030.

A Comparative Analysis of Machine Learning Models for Crop Recommendation using Explainable AI

Crop selection is a critical decision in agriculture because it directly affects yield, resource utilization, and economic sustainability. Conventional crop recommendations rely on human experience and agricultural norms, but they do not take into account variability in soil nutrients and climate factors. In this paper, a comparative analysis of machine learning algorithms for recommendation of crops based on soil and climate parameters will be done. Features such as nitrogen, phosphorus, potassium, temperature, humidity, pH, and rainfall are used in order to predict appropriate crops. In order to enhance the efficiency of the algorithm, data preprocessing, feature engineering, train–test splitting, feature scaling, and Synthetic Minority Over-sampling Technique (SMOTE) are implemented. Six machine learning models have been compared: Random Forest, XGBoost, Logistic Regression, CatBoost, Explainable Boosting Machine (EBM), and TabNet. Metrics for comparison of these models include accuracy, precision, recall, F1-score, confusion matrix, and ROC–AUC. To enhance the interpretability of the results, explainable AI methods, including SHAP, Permutation Feature Importance, and LIME, are implemented. It was found that the Random Forest model performed better than other models, and the results for XGBoost, CatBoost, and TabNet were competitive.

Self-Supervised Learning for Brain MRI Classification using Machine Learning

Brain MRI classification is essential for the early
diagnosis and treatment of brain cancers and other brain diseases
with neurological symptoms. But sometimes training an efficient
deep learning to understand from labeled MRI data is a difficult
task because of the limited availability of these data in terms of
cost effectiveness and a time consuming process. We propose a
hybrid SSL framework for brain MRI classification integrating
MRI-specific preprocessing augmentation methods with SimCLR
BYOL in this paper. The United publicly available (open-source)
brain MRI datasets were merged into a single dataset consisting
of 13604 training and testing images in four distinct classes-
glioma, meningioma, pituitary tumor and no-tumor comprising
of 4012 images.During self-supervised pretraining, the hybrid
SimCLR-BYOL framework learns complementary visual repre-
sentations through multi-view learning and anatomy-preserving
augmentations, reducing its dependence on manual annotations.
Experimental results show that the proposed SSL approach
achieves an accuracy of 96.81%, outperforming a conventional
supervised baseline that achieves 92.70% accuracy. The model
also records a macro F1-score of 96.62%, indicating reliable
performance across all categories. These findings demonstrate
that the proposed hybrid self-supervised framework can improve
brain MRI classification and provide a practical solution for
medical imaging applications where labeled data are limited

SGA-WQI: SHAP-Guided Adaptive Water Quality Index for Deep Learning-Based Water Quality Prediction Using TabNet and Transformer Models

Clean and safe water is essential for human health,
environmental sustainability, and economic development. However, water quality is rapidly degraded by industrial, agricultural,
and urban pollution. Traditional monitoring relies on manual
laboratory sampling, which is costly, time-consuming, and unsuitable for real-time analysis. To solve these limitations, this study
proposes an explainable deep learning framework for multiclass water quality classification using the Water Quality Index (WQI). Three tabular deep learning architectures—TabNet,
TabTransformer, and FT-Transformer—were evaluated. For data
preprocessing, median-based missing value imputation, normalization, and SMOTE-based class balancing were included. Model
evaluation was conducted using accuracy, precision, recall, and
F1-score, with the help of SHAP-based explainability. SHAPbased explainability helped to analyze feature contributions.
Among all models, TabNet achieved the best performance with
an accuracy of 98.3 %. This paper also introduces a novel SHAPguided Adaptive Water Quality Index (SGA-WQI) that dynamically adjusts parameter weights and captures feature interactions
which improves model interpretability and robustness for realworld environmental monitoring.

Soft Voting Ensemble with Explainable AI for Sustainable Crop Recommendation

Successful crop recommendation is critical to make sustainable agriculture and food security possible in nations where farmers largely rely on traditional methods. This study proposes a machine learning-based approach employing ensemble techniques to enhance decision-making in crop recommendation. Multiple classifiers, including Random Forest, Naive Bayes, SVM, XGBoost, and ANN, were developed and compared, with the Soft Voting Ensemble model achieving the highest accuracy of 99.55%. Further, explainable artificial intelligence techniques like LIME were used to make model predictions interpretable to end users. It is evident that machine learning, when combined with explainable techniques, offers an effective solution to support farmers with transparent and fact-based recommendations. By providing both highly accurate predictions and clear justifications for each recommendation, the system enhances farmers’ trust and confidence in adopting technology-driven solutions. This synergy between ensemble learning and explainable AI ensures reliability and transparency, making it a significant step toward data-driven agricultural decision-making.

Breaking BERT: A Comprehensive Study of Transformer Robustness to Linguistically Grounded Semantic Perturbations

The results for transformer models, such as BERT are good on many NLP benchmarks, but the models’ accuracy doesn’t guarantee understanding of the language. This paper evaluates all of these (DistilBERT, Bi-LSTM, TextCNN and Logistic Regression) with respect to paraphrase detection on the Quora Question Pairs (QQP) dataset and compares each of them against a newly created Tricky Dataset containing 51,338 perturbed pairs over four transformations: negation insertion, quantifier swap, entity replacement, and distractor prefix. DistilBERT’s top results on the standard QQP hold 89.9% accuracy, whereas it fails to get over 13.8% on the quantifier swaps, a 76.1%-point below random chance. Content-word overlap is identified as the important thing by LIME and by ”attention-map” analysis and systematic neglect of logical operators is identified in the models. Realistic expectations for the abilities of these models are higher than the scores found in the standards. A gap is revealed with a perturbation-based evaluation.

Towards Accurate Classification of Authentic Photographs and AI-Generated Images Using Deep Learning Approaches

Controlled, matched-condition comparison: five CNN backbones (ResNet18, ResNet50, VGG16, VGG19, EfficientNetB0) evaluated under a unified transfer-learning protocol on the same 120,000-image CIFAKE benchmark, isolating architecture effects from training-setup differences.
Rigorous data-integrity pipeline: corruption checks, MD5-based cross-split duplicate detection, and blur flagging applied to CIFAKE before training, closing a common data-leakage gap in prior work on this dataset.
Fine-tuning depth as the key variable: the study shows fine-tuning strategy (full vs. frozen-backbone) matters more than raw architecture choice, with fully/partially fine-tuned models outperforming frozen ones by up to ~14 accuracy points, and frozen VGG models developing a measurable class bias (many more false positives than false negatives).
A new low-data benchmark: the custom \textit{Images} dataset (2,400 images) tests whether these findings hold when only ~1,700 training images are available, addressing a real-world constraint that CIFAKE-scale benchmarks don’t capture.
Soft-voting ensemble result: combining EfficientNetB0, ResNet, and VGG on the custom dataset eliminates all remaining classification errors (Precision = Recall = F1 = 1.0000), demonstrating that ensembling can close the residual gap left by any single fine-tuned model even in a small-data regime.
Practical deployment guidance: an accuracy-vs-efficiency comparison (params, relative inference time) identifying EfficientNetB0 as the best accuracy-per-parameter option for resource-constrained, real-world content-moderation deployment.

Graph Neural Network with Graph-Conditioned Sequential Readout for Fault Localization in Power Grids

Fast bus-level fault localization is essential for
resilient power grids under renewable integration and changing
operating conditions. This work presents a Graph Neural Network with Graph-Conditioned Sequential Readout that combines
topology-aware message passing on the IEEE 39-bus system
with an LSTM readout over graph embeddings. Trained with
a multi-task objective for localization and auxiliary fault-type
prediction, the model is evaluated on a balanced synthetic dataset
of 60 544 samples. It achieves 99.8% fault-only top-1 localization,
100.0% top-3 localization, 100.0% fault detection, and 0.012 mean
absolute bus-index error. Random Forest saturates this synthetic
benchmark at 100.0%, and we identify the property of the data
generator responsible; the contribution is therefore the topologyaware formulation, the grouped evaluation protocol, and the
accompanying robustness analysis rather than a raw-accuracy
lead.

Techno-Economic Assessment of Fixed-Tilt Bifacial Agrivoltaic Systems for BRRI dhan33 Rice Cultivation in Northern Bangladesh: A Simulation-Based Approach

This study develops a simulation-based techno-economic framework for evaluating fixed-tilt bifacial agrivoltaic systems integrated with BRRI dhan33 rice cultivation in northern Bangladesh. A controlled comparison of nine location–pitch configurations is performed using hourly climate data, bifacial PV modelling, Whole-Plot crop radiation analysis, water-saving scenarios, and a 25-year financial assessment. The study identifies the trade-off between electricity generation, crop preservation, and economic performance, providing a location-specific decision framework for sustainable food-energy integration in Bangladesh.

Rice Based Agrivoltaics in Northern Bangladesh: A Structured Evidence Synthesis and a Simulation Ready Design Framework for Fixed Tilt Bifacial Systems

This paper develops a structured evidence synthesis and a simulation-ready design framework for rice-based agrivoltaic systems in northern Bangladesh. Unlike conventional approaches based only on average shading, this study emphasizes dynamic crop-level irradiance, growth-stage sensitivity, bifacial photovoltaic performance, land equivalent ratio, and techno-economic evaluation. The proposed framework integrates optical shading, PV energy modelling, crop response, water scenarios, and financial analysis to support future field calibration and sustainable food-energy planning in Bangladesh.

Analysis of Pedestrian Crossing Behavior at Unsignalized Crossings in Rome, Italy

The present study provides empirical evidence of pedestrian risk behaviour at an unsignalized mid-block crossing in Rome. It measures the level of unsafe crossing practices and shows that approximately one third of pedestrians crossed without looking in both directions and almost one quarter were distracted by mobile phones. The study also adds an age-related dimension, showing that young pedestrians are disproportionately affected by mobile phone distraction and that unsafe crossing behaviours increase significantly during peak hours.

An important result of the study is the strong impact of the group size on the pedestrian behaviour. Pedestrians in larger groups were significantly less likely to assess traffic conditions independently, showing a tendency toward herd-like behaviour. To summarise, the findings suggest that behavioural and situational factors are stronger predictors of pedestrian risk than demographic characteristics.

Pose-Based EKF-SLAM Using NDT Scan Matching and Loop Closure for Indoor Robot Navigation

Key research contributions include:
1. Robust Scan Matching: Using NDT instead of ICP for faster convergence and greater stability in sparse indoor environments.
2. Reliable Loop Closure: A rigorous validation pipeline using overlap ratio, fitness score, and Mahalanobis distance to ensure accurate loop detection and prevent false corrections.
3. Effective Sensor Fusion: Fusing encoder odometry, IMU orientation, and NDT pose corrections within a single EKF framework to maintain a globally consistent trajectory.
4. Real-World Validation: Successfully deployed on a TurtleBot in both simulation and real indoor settings, demonstrating bounded uncertainty and effective drift reduction.

Transmission Line Fault Detection System Using Flame and Moisture Sensing for Real-Time Environmental Hazard Monitoring

M. R. I. Mejba conceived and led the research, including problem formulation, system design, methodology development, data analysis, and manuscript preparation. The physical prototype was constructed by the project team under the author’s direction. All remaining work including circuit design refinement, experimental validation, results interpretation, and writing of the manuscript was carried out by the author.

SSM-Net: A Hybrid Architecture for Child Speech Emotion Recognition

This study proposes SSM-Net, a lightweight hybrid architecture for classifying child speech emotions into positive, neutral, and negative categories. Speech signals are transformed into normalized log-Mel spectrograms and processed through a compact framework that integrates a MobileNetV3-Small backbone, multi-kernel depthwise convolution, and Mamba-based state space modeling. The convolutional modules capture local time-frequency patterns, while the state space blocks model long-range contextual dependencies across the spectrogram representation. The proposed model combines MobileNetV3-Small, multi-kernel depthwise convolution, and Mamba-based sequence modeling for positive, neutral, and negative child speech emotion classification. The main contributions are –

1. A dataset for child emotion detection containing 890 samples has been accumulated and developed.
2. Developing a hybrid lightweight architecture for SER classification into 3 classes such as positive, neutral, and negative called SSM-Net that performs comparatively better than other existing lightweight models.

BanglaCLR: Structure-Aware Contrastive Representation Learning for Bengali Character Recognition

Bengali characters often differ by a short stroke, dot, or loop, making recognition sensitive to both shape and stroke layout. BanglaSAN encodes appearance and structure in separate branches, exchanges information between them through bidirectional cross-attention, and classifies the fused representation against graph-refined prototypes. Its contrastive loss assigns larger penalties to confusable negative classes. Because the 49 classes in BCRD each contain only one distinct glyph image, we generate non-overlapping degraded samples for training, validation, and testing. We compare seven model configurations and assess significance, calibration, robustness, and explanation faithfulness. BanglaSAN obtains 95.22% test accuracy and 2.9% expected calibration error. It outperforms the single-branch CNN and structure-only baselines, although several appearance-based ablations have statistically comparable accuracy. Rotation, blur, and occlusion tests favor the proposed model at high severity. Appearance-deletion experiments also show that the structure branch carries class information on its own. Gaussian noise is the main exception: corrupt edges sharply reduce the benefit of the structure pathway.

y

a

Energy-Efficient Multimodal AI for Optimizing Crop Yields in Smart Agriculture

1) We proposed Dual CMAB(Contextual Multi-Armed Bandit),a two-level bandit framework decoupling sensor-level energy management (CMAB-1) from model-level inference selection (CMAB-2) for end-to-end energy optimisation in a multimodal agriculture pipeline. CMAB-1 is a two-arm policy: Arm-0 activates only tabular sensors for irrigation and fertilizer decisions; Arm-1 additionally activates the camera for image-based disease detection.

2) We train MobileNetV2 and ResNet34 on the 9-class Paddy Doctor dataset, and build XGBoost classifiers for irrigation scheduling and fertilizer recommendation.

3) We quantify energy consumption using real GPU measurements on an NVIDIA GeForce RTX 5060 Ti (mil-
lijoules per sample), demonstrating 23.1 % real energy savings versus always ResNet34 and 88.6 % simulated
savings versus a static always-on pipeline, with negligible accuracy degradation.

CONFIRM: Comprehensive, Orchestrated, Non-Negotiable Framework for Information Security and Risk Management

This research contributes “CONFIRM (Comprehensive Orchestrated, Non-Negotiable Governance Framework for Information Security and Risks Management)” as an integrated governance framework designed to eliminate information-security governance blind spaces through continuous end to end visibility and traceability. The framework establishes a traceable relationship from “Organizational Vision → Mission → Objectives → Goals → Activities → Evidence/Status”, while integrating “people, processes, technologies, enterprise assets, risks, controls, compliance obligations, security operations, and resilience” within a unified governance architecture. Its principal contribution is the transformation of fragmented operational and security information into continuously traceable governance intelligence through “Enterprise Relationship Intelligence (ERI)” and a “Continuous Governance Intelligence (CGI)” cycle. Unlike conventional point-in-time or siloed governance approaches, CONFIRM enables responsible and accountable parties, management, the Board, and relevant authorities to determine the current status, ownership, dependencies, risks, control effectiveness, evidence, and business impact of governance activities and to trace these upward to organizational objectives. The framework further introduces “automated multi layer processing, quantitative governance/risk/resilience measurement, and AI embedded decision support”, enabling early identification of deviations, evidence based decisions, timely corrective actions, and measurable assessment of whether information security activities along with other operational and business activities are effectively supporting organizational objectives and the achievement of the organizational Vision.

Privacy-Preserving Single-Lead Smartwatch ECG Monitoring for Multi-Class Arrhythmia Detection

This work presents a fully local, privacy-preserving inference framework, ECGNet, for multi-class arrhythmia detection from simulated 1-lead smartwatch ECG, extending beyond binary Normal/AFib detection to four clinically relevant rhythm classes. A confidence-calibration mechanism withholds low-confidence predictions rather than forcing unreliable diagnoses, and a fair 12-lead vs. single-lead benchmark under an identical model quantifies the diagnostic cost of wearable-grade acquisition, achieving 90.64% accuracy with a 9.36% inconclusive rate at a calibrated 0.80 confidence threshold. A hardware-agnostic sampling-rate simulator further evaluates robustness across device-specific acquisition rates without requiring physical wearable hardware.

Multi-Backbone Ensemble Learning for Robust Breast Cancer Classification on the Mini-DDSM Mammography Dataset

This study presents a rigorous patient-grouped evaluation framework for breast cancer classification on the Mini-DDSM dataset, while identifying and addressing cross-folder patient-identifier collisions that can distort patient grouping and fold composition. It systematically evaluates EfficientNet-B4 and ConvNeXtV2-Tiny using test-time augmentation, threshold analysis, and Grad-CAM explainability, providing a reproducible reference framework for robust mammography classification.

OTSM-Configured Strategic Communications for Integrated 6G Autonomous Metro Rail Networks

The paper introduces a strategic communication backbone tailored specifically for integrated 6G autonomous metro rail networks. By modeling the complex integration of wireless sensors directly into the metro rail’s core communication framework, the research provides a targeted architectural solution designed to minimize latency in mission-critical, real-time metro communications.

Fabrication and Performance Analysis of a Zinc-Bromine Single Flow Battery

A low-cost single-flow zinc-bromine cell is fabricated in which the positive-side reservoir and pump are eliminated, using non-woven glass fiber as the separator – a material established in lithium-based and lead-acid systems together with a graphite positive electrode backed by porous carbon felt to increase active surface area and retain evolved bromine.

Off-Grid Electrification of Urir Char Using HOMER-Optimized Hybrid Renewable Microgrids

Addressing the critical challenge of off-grid island electrification in Bangladesh, this study presents the first comprehensive HOMER Pro techno-economic and environmental evaluation of a hybrid renewable microgrid tailored for Urir Char. The primary contribution lies in the novel integration of solar PV, hydrokinetic turbines harnessing strong estuarine water currents, biogas generators utilizing local agricultural and animal waste, and Lithium Iron Phosphate battery storage, a system architecture previously unexamined in HOMER Pro literature for Bangladeshi off-grid coastal regions. The study demonstrates that an optimal 100% renewable Load Following configuration (PV/Biogas/Hydrokinetic/Battery) delivers reliable energy access to 16,078 residents at an exceptionally low Levelized Cost of Energy (LCOE) of $0.0629/kWh and a Net Present Cost (NPC) of $3,134,828. Furthermore, the findings prove that a locally fueled biogas generator operating for just 353 hours per year effectively eliminates the high economic, logistical, and environmental risks associated with diesel transport while maintaining near-zero carbon emissions (507 kg/year), highlighting that dispatchable bio-energy, rather than over-sizing solar or battery capacity, is essential to resolving renewable intermittency.

Effects of Categorical Encoding, Normalization, and Feature Reduction on Ensemble Learning for Mushroom Edibility Classification

Mushroom edibility classification is an important problem because poisonous mushrooms may cause serious health risks. Machine learning methods can be useful for classifying mushrooms using their different characteristics. However, preprocessing methods and the number of features can affect classification performance. In this study, the effects of categorical encoding, normalization, and feature reduction are investigated using the Secondary Mushroom dataset. Two categorical encoding methods, label encoding and frequency encoding, are considered. Min-max normalization is also applied to examine its effect on the results. Four ensemble machine learning models, namely Random Forest, XGBoost, LightGBM, and AdaBoost, are evaluated. The features are ranked according to their importance, and the models are tested using the top 2 to 20 features. The experiments use an 80:20 hold-out split with a fixed random state. The results show that frequency encoding improves the performance of XGBoost and LightGBM for different feature subsets. Normalization produces almost no change in the accuracy and feature ranking of the investigated tree-based models. Random Forest and LightGBM achieve at least 99.9% accuracy using only eight features. The results indicate that categorical encoding and feature reduction can affect the performance and feature requirement of ensemble models for mushroom classification.

LiteDR-XAI: A Lightweight and Explainable Framework for Referable Diabetic Retinopathy Screening

Automated diabetic retinopathy screening must balance predictive performance with computational cost, threshold control, reproducibility, and interpretability. This paper presents LiteDR-XAI as a reproducible framework for binary referable diabetic retinopathy screening on APTOS 2019, not as a new con volutional architecture or a universal state-of-the-art model. Five pretrained backbones are evaluated through 35 controlled runs covering input resolution, contrast-limited adaptive histogram equalization, pretraining, backbone freezing, and loss design. Checkpoints are selected by validation AUROC, and the operating threshold is estimated only from validation predictions using Youden’s index before application to the held-out test set. From these experiments, a compact MobileNetV3-Large configuration at 320 pixels is selected as the LiteDR-XAI operating model because it provides a balanced accuracy-efficiency-explainability trade-off: 94.73% accuracy, 0.9791 AUROC, 0.9646 AUPRC, 95.96% sensitivity, 93.88% specificity, 91.45% precision, and 93.65% F1-score with 3.22 million parameters. A strict literature comparison includes only near-direct APTOS binary/referable DR studies and shows that LiteDR-XAI has the highest accuracy among the included comparisons, while other methods report higher AUC or sensitivity under different protocols. Grad CAM supports qualitative inspection of correct and erroneous predictions. External, multi-seed, device-level, and prospective validation remain necessary before deployment.

A Rule-based Automatic Algorithm for Detecting Total Sleep Time using Pulse Oximetry Signal

This study proposed a lightweight and interpretable rule-based framework to estimate TST from overnight finger-tip pulse oximetry signals only. After pre-processing, different features were extracted from each 30 second epoch and the rule-based framework classified each epoch as either sleep or wake. After applying TST correction, the framework finally estimated the TST as the output. Performance evaluation demonstrated significant correlation between estimated and annotated TST values (Pearson’s correlation coefficient, r = 0.804 and Lin’s concordance correlation coefficient, CCC = 0.760). Bland-Altman plot demonstrated small mean bias (0.24 hours) with a narrow limits of agreement (-0.95 to 1.44 hours). The proposed framework provides reliable and computationally efficient TST estimation, thus a promising approach for portable and home-based sleep monitoring applications.

Technological Innovation in Precision Agriculture: An EfficientNet-B0 Transfer Learning Framework for Grape Leaf Disease Diagnosis

This paper advances mobile precision agriculture by engineering an optimized EfficientNet-B0 transfer learning framework that resolves the critical trade-off between high diagnostic precision and edge-compute efficiency in vineyard biosecurity. Unlike legacy architectures like VGG16 that suffer from high parameter redundancy (138M+ parameters) and slow inference latency, our compound-scaled model achieves a peak 94.80% diagnostic accuracy and 94.80% F1-score across field-collected grape foliage classes while maintaining a remarkably lightweight profile of 5.3 million parameters, a 21 MB memory footprint, and an execution cost of just 0.39 GFLOPs. Furthermore, by evaluating performance directly on outdoor, variable-field imagery rather than controlled laboratory benchmarks, this work demonstrates a practical, parameter-efficient pipeline tailored for real-time, offline edge-AI deployment on resource-constrained mobile hardware and autonomous agricultural drones.

TRACE-BN: Transferring Bangla-English Tutoring Behavior to a Sub-1B Offline Language Model

This paper introduces TRACE-BN, a curriculum-guided bilingual tutoring dataset of structured traces for Bangla-to-English language learners at the CEFR A1–A2 level. Each trace encodes a complete seven-field pedagogical sequence combining word-level glosses, literal and natural translations, contrastive Bangla grammar explanations, common learner mistakes, and targeted practice with answers. We show that a compact sub-1B offline model (Qwen3-0.6B) can acquire this multi-component tutoring behavior via 4-bit LoRA fine-tuning, improving schema validity from 85.4% to 95.8% while raising translation and pedagogical scores. We validate the approach through an expert human audit of the supervision signal and a reference-aware dual-judge evaluation showing consistent improvements across all evaluated tutoring dimensions.

SafeNet-Guard: A Secure Reporting Platform for Technology-Facilitated Harassment with Evidence Integrity and Duress Signaling

Four contributions follow from our proposed design: (i) a role-stratified secure reporting architecture, (ii) tamper-evident versioning of reports and evidence, (iii) a dual-credential coercion-response protocol, and (iv) a controlled evaluation that stress-tests these components against tampering and coercion scenarios.

First-Principles Study of Lead-Free Cs2NaGaBr6-xFx (x = 0.00, 0.22, 0.33) Perovskites for Next-Generation Functional Materials

In this work, a systematic density functional theory study on lead-free double perovskites Cs₂NaGaBr₆₋ₓFₓ (x = 0.00, 0.22, 0.33) is reported, which reveals the changes brought by F substitution on the structure, electronic, optical, mechanical and elastic properties of these compounds. This paper finds out the variations in electronic structure and band gap parameters along with the enhancement in optical response and absorption of these compounds, which suggests that halide engineering can play an important role in tuning the functionalities of these materials. The discussion on the elastic and mechanical properties of the materials helps in understanding the structural stability, mechanical strength for application of next generation materials.

Efficient Network Intrusion Detection Using Random Forest and Lightweight REST API Deployment

Comparative Techno-Economic Analysis of Si, GaAs, and InGaN Thin-Film Solar Cells under Bangladesh’s Climatic Conditions

The performance of photovoltaic (PV) systems is influenced by operating temperature, especially in Bangladesh’s tropical climate. This study presents a comparative numerical investigation of Si, GaAs, and InGaN p-i-n thin-film solar cells using site-specific meteorological data. Temperature-dependent photovoltaic characteristics were investigated using SCAPS-1D simulation software. Quantum efficiency (QE) and spectral response (SR) were evaluated to characterize optical absorption performance across the AM1.5G solar spectrum. GaAs achieves the highest power conversion efficiency (PCE) of 23.61% due to its superior thermal stability. InGaN demonstrates potential for high-temperature environments, yielding the highest open-circuit voltage and a stable fill factor due to its wide tunable bandgap. In contrast, silicon exhibited the highest thermal sensitivity and the lowest PCE (10.51%) due to the lower light absorption of the thin absorber layer. Furthermore, economic feasibility assessment using RETScreen for a proposed 10 MW utility-scale power plant revealed that GaAs-based system requires the smallest solar collector area (41,391 m²) and generates 16,530.91 MWh annually at the lowest cost of electricity ($0.0777/kWh). In contrast, the silicon-based system requires the largest collector area and incurs the highest energy production cost.

CyberMirror: Recovering Interpretable Attacker Reward Functions from Deep-RL Pentesting Agents via Inverse Reinforcement Learning

1. A demonstration-free pipeline on an emulation-friendly
range that transforms a PPO/MLP attacker into an
attacker-value detector without any human data, using
a sample-based Maximum-Entropy IRL reward recovery
algorithm with a separable objective and real partition
function, made to converge by (i) Markovian state-
occupancy features, (ii) horizon-matching the background
to the median demonstration length, and (iii) Ridge
regularization (λ = 0.3).

2. Empirical evidence that the recovered score differentiates
adversarial from benign states (AUC ≈ 1.0, 100% detect-
before-goal), and a detection-gated, shared-environment
duel in which this reward incentivizes an action-taking
defender, reported with its availability cost so that con-
tainment is not mixed up with free defense.

3. A transferable methodological finding on the trade-off be-
tween attack optimality and reward identifiability in IRL-
from-RL, explicitly stating the limitations and providing
a future roadmap.

Exploring SVM and CNN Behavior Across Kepler Optical Light Curves and TART Radio Visibilities

The research study is organized around the following objectives and contributions:
• Formulate a shared binary quality-control experiment that distinguishes original observations from deterministically corrupted copies using four analogous, modality-specific corruption categories at three severity levels.
• Construct provenance-tracked, source-disjoint Kepler and TART workflows with an equal-condition stress grid applied to every held-out source.
• Explore SVM and CNN behavior through a matched cross-modality design using balanced test metrics, source-level bootstrap intervals, and a 60-session TART sensitivity analysis, without claiming novelty for the individual algorithms or corruption concepts.

Operational Assessment of a Small-Scale Solar PV System : A Case Study from Bangladesh

The goals are to: (i) validate
the consistency of the solar datasets; (ii) quantify the daily/period
yield, FLH, specific yield and ca-capacity factor; (iii) test the
dataset sensitivity with six categories of the logged weather
data; and (iv) position the results in the context of the regulatory
framework of SREDA and other similar, peer-reviewed regional
studies

Social Media Sentiment Analysis for Local Businesses

This study develops a unified Yelp-based sentiment analysis framework that systematically compares lexicon-based, traditional machine-learning, and transformer-based models under a common evaluation setting. The best-performing model, DistilBERT, is then applied to a separate 300,000-review sample to support category-, aspect-, and business-level analysis. The main contribution is the integration of sentiment classification with keyword-based aspect detection and category-relative benchmarking to identify interpretable business strengths, weaknesses, and actionable customer-feedback patterns.

Explainable and Fair Student Dropout Prediction from Multimodal Temporal Learning Analytics

We propose a multimodal temporal framework that integrates demographic, academic, and week-by-week VLE behavioral data for early student dropout prediction.
We develop a dual-input BiLSTM architecture that jointly models static student characteristics and temporal learning engagement patterns to improve predictive performance.
We incorporate SHAP and LIME to provide both global and student-level explanations, enabling educators to understand the key factors underlying dropout predictions.
We conduct a systematic fairness assessment across gender, age, disability status, and socio-economic deprivation to identify potential disparities in model performance and prediction errors.
We translate predictive and explainability outputs into risk levels and personalized intervention recommendations, supporting actionable early-warning decision-making in virtual learning environments.

Real-Time Blind Spot Vehicle Detection and Risk Classification System Using Deep Neural Network

This study is about a way to detect things around autonomous vehicles in real time. It uses a two step process. First it finds objects using YOLOv8. Then it figures out what those objects are using other methods like EfficientNetV2, MobileNetV2, VGG16 and ResNet-50. The study used a lot of pictures of roads around 9,850 of them to make this system work. EfficientNetV2 was the best at getting it right it was correct 91 percent of the time and it was very good at finding things that could be dangerous. This system is as good as or even better than systems that use radar. It only needs one low cost camera. The study also looked at how these methods work for detecting blind spots, which is a big problem on highways in Bangladesh where there are a lot of different types of vehicles on the road and accidents happen because people cannot see what is, around them.

An Empirical Comparison of Monolithic and Microservices Architectures for an E-Commerce Application

This paper presents an empirical performance comparison of monolithic and microservices architectures for an e-commerce application. Using k6 load testing at 50 and 100 virtual users, we demonstrate that microservices achieve 5.4% higher throughput, 25% lower average latency, and 39% lower p95 latency at 100 VUs, while exhibiting different failure modes. The monolith shows consistent order-creation failures (~0.7% error rate), whereas microservices failures are transient and service-specific. This study provides quantitative evidence that architectural decomposition can improve scalability and tail latency under stress, with implications for software architecture decision-making.

A Comparative Analysis of Multimodal Frameworks for Multilingual Human vs. AI Caption Detection

This paper presents a comprehensive comparative analysis of multimodal frameworks for detecting human vs. AI-generated captions in a multilingual context. It identifies the most robust framework architectures and provides critical insights into cross-lingual detection accuracy, helping improve the reliability of AI-generated content auditing.

MedlmBD: An AI-Driven Framework for Bengali Healthcare Query Management

• Designing a medical AI assistant that addresses key gaps in existing research
• Developing and fine-tuning the SmolLM3-3B model using a subset of the ChatDoctor Dataset
• Performing a comprehensive evaluation and comparison with a pre-trained medical domain model for evaluating the quantitative and qualitative metrics of our architecture
• Overcoming language barriers in healthcare support through the use of a Bengali-English translation pipeline.
• Despite using only a fraction of the parameters and being fine-tuned only on a subset of the ChatDoctor dataset, fine-tuned SmolLM3-3B is competitive with the larger MedAlpaca-7b model.

CropGuard AI: A Crop Disease Diagnosis System Combining EfficientNet-B0 and Vision-Language Models for Bangladeshi Agriculture

The research contributes an integrated crop disease diagnosis framework that combines EfficientNet-B0 image classification with vision-language diagnosis, Bangladesh-focused disease knowledge, and RAG-based agricultural advisory to provide more comprehensive and accessible crop disease support.

A Leakage-Aware Ordinal Machine Learning and Explainable AI Framework for Hidden Replication, Label Ambiguity, and Performance Inflation in Maternal Health Risk Prediction

This paper proposes four significant contributions:
1) We quantify the scale of duplication and label ambiguity and its practical consequence for standard train/test splitting.
2) We directly measure the resulting performance inflation by comparing conventional cross-validation against exact-row-grouped and physiological-profile-grouped cross-validation for two widely used models.
3) We benchmark a broad family of ordinal-aware and conventional classifiers under leakage-aware, repeated nested cross-validation, including a two-threshold ordinal decomposition and a CORAL-style ordinal neural network, with full calibration and conformal-prediction analysis.
4) We evaluate the stability of SHAP-based explanations across random seeds, treating explanation reliability itself as an empirical outcome rather than an illustrative afterthought.

An Explainable and Robust Machine Learning Framework for IoT Intrusion Detection through Zero-Day and Cross-Dataset Evaluation

This research proposes a comprehensive and repeatable machine learning system for IoT IDS, covering leakage-aware pre-processing, comparison of five ML models, class imbalance, explainability using SHAP and a lightweight 15 feature IDS. The framework also assesses the robustness to zero day attacks under a Leave-One-Attack-Out approach and compares the supervised detection approach with an Isolation Forest baseline. Furthermore, bidirectional cross-dataset evaluation between TON_IoT and UNSW_NB15 is conducted to evaluate the model in different network environments. The results show that explainability, resistance to unseen attacks, and cross-dataset validation are necessary to measure the usability of intrusion detection systems in IoT.

A Hybrid Multi-Backbone Deep Learning Framework with Learnable Attention Fusion for Explainable Potato Leaf Disease Detection

Potato leaf diseases, particularly early blight and late blight, significantly reduce crop yield and quality, making accurate and timely diagnosis essential. This paper introduces a hybrid multi-backbone deep learning framework, which combines EfficientNet-V2-S, Swin Transformer and ViT-Base, and introduces a Learnable Attention Fusion (LAF) module to capture complementary local and global feature representations. A progressive layer-freezing approach is used to enhance transfer learning and model generalization. The proposed framework is tested in terms of 5-fold cross validation and cross dataset testing. The experimental results are obtained with good average classification accuracy of 99.47% for 5-fold cross validation and 97.34% for cross-dataset evaluation, exhibiting good robustness and generalization capability for different data distributions. Moreover, Grad-CAM is used to give visual interpretation of model predictions by identifying the relevant areas of the image associated with the disease. The proposed model is used as a web application for practical agricultural applications, allowing identification of potato leaf diseases in real-time based on uploaded leaf images. From the experimental results, it is proved that the proposed framework gives an accurate, robust and interpretable solution for the potato leaf disease automatic classification.

Multi-Class Brain Tumor MRI Classification Using EfficientNet Based Hybrid Attention Network

Brain tumor diagnosis from magnetic resonance imaging (MRI) plays an important role in supporting early clinical assessment. However, distinguishing different tumor types remains challenging because of variations in tumor size, shape, and appearance across patients. This paper presents an enhanced EfficientNetV2B0-based model that combines the Convolutional Block Attention Module (CBAM) with Multi-Head Self-Attention (MHSA) to improve feature representation for multi-class brain tumor classification. The model is evaluated on a publicly available dataset containing 7,200 brain MRI images from four categories: glioma, meningioma, pituitary tumor, and no tumor. Its performance is compared with EfficientNetV2B0, EfficientNetV2B0 with CBAM, EfficientNetV2B0 with MHSA, ResNet50, DenseNet121, and MobileNetV3Small under the same experi- mental setting. The proposed model achieved a test accuracy of 96.30%, with a weighted F1-score of 96.29%, a mean Dice score of 96.36%, and a mean IoU of 93.02%, outperforming all baseline models. It also obtained macro and micro AUC values of 99.65% and 99.63%, respectively, demonstrating consistent clas- sification performance across all classes. These findings indicate that integrating channel attention with self-attention helps the network learn more discriminative features while preserving the computational efficiency of the EfficientNetV2B0 backbone.

Renewable Energy Consumption and Sustainable Wealth in Developing Countries: The Moderating Role of Government Effectiveness

These outcomes highlight the importance of strengthening institutional quality alongside renewable energy expansion to promote sustainable wealth accumulation in developing countries.

Graph-Based Clustering and Explainable Graph Neural Networks for Crime Hotspot Prediction in Bangladesh (2020–2025)

Bangladesh has experienced noticeable variation in crime across its metropolitan and range-level policing units during 2020–2025. Some regions, such as Dhaka and Chattogram, regularly report high volumes of violent and property crimes, while others show more sporadic spikes. Conventional analysis tools, for example, trend plots or static heatmaps, often describe these patterns in isolation and struggle to capture how crime in one area resembles or relates to crime in another. This paper presents a graph-based framework that combines community detection, Graph Neural Networks (GNNs), traditional machine learning, and explainable AI (XAI) to anticipate high-risk regions and to understand how policing units are connected in terms of crime behaviour. Monthly crime data from 17 units are used to construct a similarity graph where nodes represent units and edges represent similarity in crime statistics. Louvain and spectral clustering reveal three stable crime-behaviour communities, and the same engineered features feed Random Forest, Graph Convolutional Network (GCN), and Graph Attention Network (GAT) models for next-month high-risk prediction. Random Forest achieves the highest overall accuracy, while the GAT model attains the highest recall on high-risk months, making it useful where missing an emerging hotspot is costlier than a false alarm. SHAP and GNN Explainer are applied to interpret the Random Forest and GCN respectively, showing that Random Forest concentrates its decisions on a small set of temporal features while the GCN distributes importance more evenly across features and relies on a specific subset of high-weight connections between policing units (importance-ranking correlation between the two methods = 0.5125). The results indicate that graph-based analysis, combined with explainability tools, can make crime forecasts more interpretable and useful for planning, even when a classic tabular model remains the most accurate overall predictor.

A Machine Learning Approach to Screening for Child Undernutrition in Rural Bangladesh

This paper builds separate stunting, underweight, and wasting classifiers for rural under-five children from BDHS 2022, correcting class imbalance strictly inside a leakage-free training pipeline (SMOTE and class weighting applied only after the train–test split) and using recall as the primary metric. A class-weighted logistic regression gives the best recall on all three indicators and detects roughly twice as many wasted and underweight children as the closest published model on the same survey. A high-accuracy random forest that identifies none of the wasted children is shown to demonstrate why recall must be prioritised over accuracy on this imbalanced data.

Reference-Conditioned Healthy-Like EEG Reconstruction Using Movement-Aware CycleGAN for Motor Imagery BCI

Proposes a movement-aware, reference-conditioned CycleGAN to reconstruct healthy-like EEG from stroke motor-imagery EEG using movement-matched healthy references and anti-copy and movement-consistency constraints, with subject-level holdout evaluation to reduce data leakage and separately assess healthy-domain similarity and movement preservation.

Domain-Aware Hybrid Quantum Learning via Correlation-Guided Circuit Design for Crime Pattern Analytics

– Developed a comprehensive quantum-classical comparison for crime analytics with statistical validation through cross-validation across quantum, classical, and hybrid paradigms.
– Proposed a novel quantum circuit architecture that utilizes crime feature correlations through targeted entanglement based on Spearman correlation analysis. — Implemented hybrid integration strategies, including Q→C (quantum feature extraction followed by classical classification) and C→Q (classical dimensionality reduction followed by quantum modeling).

Addressing Bandwidth and Isolation Challenges in THz MIMO: A Dual-Band Bow-Tie Antenna for Biomedical, Pharmaceutical, and Security Applications

Terahertz (THz) MIMO antenna systems offer strong potential for next-generation sensing, imaging, and wireless communication, but practical deployment is still limited by narrow impedance bandwidth and poor inter-element isolation in compact designs. This work presents a dual-element MIMO antenna on a quartz substrate (170 × 68 µm²) that addresses both challenges using a modified bow-tie radiator with embedded rectangular slots. The slot-loading creates independent resonant paths within a single element, enabling dual-band operation at 5.12 THz and 6.78 THz without increasing footprint. Full-wave simulations (CST Microwave Studio) show a combined −10 dB bandwidth of 757 GHz, with return losses of −40.64 dB and −35.5 dB and VSWR values of 1.01 and 1.03, indicating near-ideal impedance matching. The antenna achieves peak directivities of 6.22 dBi and 7.44 dBi with stable linear polarization. Inter-element isolation remains better than −30 dB across the band without any decoupling structures. MIMO performance metrics further confirm strong diversity behavior, with an ECC of 1.25 × 10⁻⁷, CCL of 6.84 × 10⁻¹⁵ bps/Hz, and diversity gain of 9.99 dB, close to the theoretical limit. The proposed design serves as a compact, fabrication-compatible platform for biomedical imaging, pharmaceutical spectroscopy, security screening, and 6G MIMO communication within a single THz antenna architecture.

Design and Analysis of LP01 Mode Hollow-Core Anti-Resonant Fiber with Hybrid Elliptical Geometry

This research endeavor represents a novel air filled anti-resonant hollow-core fiber (HC-ARF) designed to achieve ultra-low transmission loss as well as efficient LP₀₁ (single-mode) guidance. The fiber geometry is fine-tuned through successive parameter sweeps. This reduces the interaction between the guided optical field and the surrounding silica significantly, which lowers the propagation loss and improves the transmission performance. The optical characteristics of the proposed fiber are analyzed using FEM (Finite Element Method). The simulation results demonstrate an ultra-low confinement loss of 4.16 × 10⁻⁶ dB/m over the wavelength range of 1.4–1.7 μm. In addition, the fiber attains a Higher Order Mode Extinction Ratio (HOMER) approaching 100 which indicates the effective suppression of higher order modes and stable single mode operation. The bending performance is also excellent, with a value of only 7.35 × 10⁻⁵ dB/m at the standard telecommunication wavelength of 1.55 μm for bending radii larger than 10 cm. These results demonstrate that the proposed HC-ARF achieves low loss and high mode purity while retaining good bending resistance, all within a simple structural design. Its low loss and modal stability suit it to long haul optical communication and high speed data transmission, whereas the hollow core itself opens further use in optical sensing and high-power laser delivery around 1.55 μm.

High-Efficiency Lead-Free Cs2PdBr6 Perovskite Solar Cells: A SCAPS-1D Numerical Study of Transport Layer Optimization

In this study, a lead-free Cs2PdBr6-based per-ovskite solar cell (PSC) has been numerically simulated systematically with help of SCAPS-1D software with an FTO/ETL/Cs2PdBr6/HTL/Au architecture. Three electron trans-port materials (WO3, TiO2, and SnS2) and three hole transport materials (MoO3, CuI, and GO) were comparatively evaluated across varying layer thicknesses. The optimum thicknesses of the FTO and absorber layers are determines to be 0.01 μm and 5 μm, respectively. Among the investigated transport materials, WO3 and MoO3 used as the electron transport layer (ETL) and the hole transport layer (HTL), respectively, due to their favorable energy band alignment and efficient charge carrier extraction capability. The combined effect of absorber defect density (1010 to 1020 cm−3) and operating temperature (280 K to 360 K) was further analyzed, confirming that crystal quality is the dominant performance-limiting factor. The fully optimized
FTO/WO3/Cs2PdBr6/MoO3/Au device achieves a PCE of 30.62%,Voc = 1.30 V, Jsc = 27.29 mA/cm2, and FF = 86.31% at 300 K, sur-passing all previously reported Cs2PdBr6-based configurations.

MediLink: Design and Validation of a Portable IoT System for Physiological Monitoring

The major contribution of this research is the development of MediLink, a portable ESP32-S3-based IoT platform for simultaneous monitoring of HR, SpO₂, temperature, and single-lead ECG. The system integrates local visualization, GPS tracking, and cloud-based remote monitoring using a FreeRTOS-based architecture. The study also provides a preliminary evaluation against reference devices and discusses the limitations of the validation process.

CALIBER: Pre-Generation Gated Image-Exemplar-Backed Retrieval-Augmented Generation for Trustworthy Dermoscopic Decision Support

The significant research contribution is the CALIBER framework, a calibrated visual-exemplar retrieval-augmented pipeline for dermoscopic decision support that combines DermCLIP-based retrieval, evidence sufficiency checking, conformal prediction, diagnostic confidence thresholds, and abstention. It improves not only diagnostic accuracy but also calibration, faithfulness, and evidence-grounded clinical reasoning, achieving 89.2% accuracy and demonstrating more trustworthy VLM-assisted dermatology decisions.

Performance Evaluation of a DeMod-CDSC-Based P-Class PMU Under Noisy Conditions

The key contribution is the first systematic noise-robustness assessment of the frequency-adaptive DeMod-CDSC P-class PMU framework, demonstrating superior TVE and FE performance over SOGI-CDSC across 30–80 dB SNR and establishing its suitability for noisy real-world smart-grid applications.

Beyond Technical Design: Assessing Stakeholder Readiness and Adoption Barriers for Blockchain-Based Traceability in Bangladesh’s Pharmaceutical Supply Chain

This study contributes by shifting the focus of pharmaceutical blockchain research in Bangladesh from technical system design to stakeholder readiness and adoption barriers. Using the Technology–Organization–Environment (TOE) framework, it examines the technological, organizational, and regulatory factors influencing blockchain-based traceability adoption among manufacturers, distributors, pharmacies, and regulators. The study proposes a context-sensitive readiness framework that highlights the roles of digital capability, implementation cost, trust, inter-organizational collaboration, and regulatory support, providing a foundation for future empirical research and practical policy initiatives in Bangladesh’s pharmaceutical supply chain.

Factors Influencing Blockchain Technology Adoption Among Small and Medium Enterprises (SMEs) in Bangladesh’s Supply Chains: A Technology-Organization-Environment (TOE) Perspective

This study addresses an important research gap by focusing on blockchain adoption among SMEs in Bangladesh’s supply chains, a context that has received limited attention in prior research. It synthesizes findings from Bangladesh and other developing economies using the Technology–Organization–Environment (TOE) framework and Diffusion of Innovations (DOI) theory to develop a context-sensitive blockchain adoption framework. The proposed framework highlights technological, organizational, environmental, and institutional factors relevant to Bangladeshi SMEs and provides a foundation for future empirical research and policy discussions.

TCA-Net: A Tri-Stage Cross-Attention Network with Explainability-Consistency and Boundary-Aware Learning for Generalizable Medical Image Segmentation

Existing medical image segmentation methods often address cross-site generalization, boundary accuracy, and explainability independently, leaving a gap in their unified integration. To address this, we propose TCA-Net, a novel CNN–Transformer framework that combines four-level Tri-Stage Cross-Attention fusion with training-time XAI-Consistency learning and boundary-aware supervision. Unlike conventional feature-fusion methods, TCA-Net explicitly captures complementary channel, spatial, and cross-attention information while promoting anatomically meaningful and precise segmentation. The framework is validated using leakage-free cross-site evaluation and controlled ablation studies, achieving 96.28% Dice in-domain and up to 98.29% Dice across sites.

IoT-Powered Hydroponic Tomato Monitoring System

This thesis demonstrates innovative use of IoT technology in hydroponic tomato cultivation, enabling real-time monitoring and control, which significantly advances precision agriculture.

Performance Evaluation of Contrast Enhancement Techniques for YOLOv8-Based Fall Detection under Low-Light Conditions

The detection of falls in low-light indoor environments is a challenge for vision-based systems because of the low visibility of features and degradation of contrast. Deep learning techniques with the YOLO algorithm provide an alternative to contact-based measurements but struggle under poor lighting conditions, and existing solutions are based on expensive IR or thermal cameras. In this study, a structured evaluation framework is proposed to systematically study the effect of contrast enhancement on fall detection with conventional RGB cameras in a controlled low-light indoor environment. Three enhancement techniques, namely Histogram Equalization (HE), Adaptive Equalization (AE), and Contrast Stretching (CS), are tested under various lighting conditions. The results show the consistent and systematic relationship between contrast restoration and the detection performance, with global enhancement showing more stable and precise results. Higher accuracy is obtained in darker scenes (75%) and dimly lit scenes (62.5%), with a precision of 1.000 in darker scenes with a high level of accuracy and thus reducing false alarms. The proposed framework can generate reproducible evidence for improving low-light detection performance and does not need any special imaging hardware or increase the complexity of models, which can promote the development of stable and scalable indoor fall detection systems for real applications.

Context-Aware E-Commerce Recommendation via Knowledge Graph- Enriched Graph Neural Networks

We chose a knowledge-graph-plus-GNN approach
specifically because it addresses this gap without abandoning
collaborative signals altogether. Rather than replacing user-
item interaction data with content-based rules, the proposed
pipeline keeps the interaction graph as the backbone and
augments it with structured attribute edges (Item–Brand,
Item–Category), allowing the GNN to propagate preference
signal along attribute pathways that a plain bipartite model
cannot see. Neo4j was chosen as the underlying store because
it allows this structure to be expressed, queried, and inspected
directly as a graph via Cypher and the Graph Data Science
library rather than reconstructed implicitly inside a black-box
model, which also makes the resulting recommendations
easier to explain to a non-technical stakeholder (Section V-E
gives a concrete example of this). Critically, we did not
simply assume that adding this context would help; the
ablation design in Section V-E was built specifically to test
that assumption and report an honest, attributable answer
rather than an unverified claim.

A Cyber Threat Intelligence Knowledge Graph for Exploit Link Prediction and Asset Risk Scoring

Existing CTI knowledge graph work stops at construction, extracting entities and relations from text but never reasoning over the finished graph. This study closes that gap by building a graph from structured feeds (MITRE ATT&CK, NVD, CISA KEV) and adding two reasoning layers: RotatE-based exploit link prediction, which roughly doubles the TransE baseline (MRR 0.173 vs 0.077), and an explainable graph-proximity risk score that ranks internal assets down to the CVE and actor behind each score. It also benchmarks LLM-based graph reconstruction against the same graph, showing structured feeds and text extraction are complementary. The core contribution is turning a static CTI graph into a tool that answers the questions analysts actually ask.

Personalized Thyroid Disorder Diagnosis System through Adaptive Expert Routing and Intelligent Selection

The main contribution of this work is AERIS, an adaptive expert-routing framework that performs patient-specific selection among heterogeneous machine learning models for thyroid disorder diagnosis. By combining a meta-learning router with Random Forest, LightGBM, CatBoost, and XGBoost, the framework achieves 97.87% accuracy and 95.98% F1-score, while SHAP-based analysis provides interpretable insights into the clinical features influencing predictions.

TriYield-XAI: An Explainable Ensemble model for Weather and Geospatial-Driven Aman, Aus, and Boro Yield Prediction in Bangladesh

We developed TriYield-XAI, a unified explainable framework for jointly predicting Aman, Aus, and Boro rice yields by integrating climatic, geographical, temporal, station-level, and historical yield information. We introduced a panel-aware hybrid ensemble architecture that combines Profile-KNN, matrix reconstruction, out-of-fold ensemble learning, meta-stacking, calibration, and residual correction to effectively capture complex spatio-temporal yield patterns and improve prediction performance. Furthermore, we applied SHAP-based explainable AI techniques to investigate the contribution of temporal, geographical, climatic, and station-specific factors to rice yield predictions, enhancing the interpretability and transparency of the proposed model.

A Fully Offline BLE-Connected Multi-Task TinyML Weather Station with Android Monitoring

The study’s main contribution is the integration of calibrated ultrasonic rainfall sensing, a three-output multi-task TinyML model on ESP32-S3, BLE communication, and an Android monitoring application into one fully offline weather station. This enables local next-day rainfall, humidity, and rainfall-risk prediction without relying on Wi-Fi, Internet, or cloud services.

Multi-Class Chest Disease Classification from X-Rays Using PCA-Enhanced Vision Transformers and Multi-Method Explainability

Main Research Contributions:
– Dimensionality reduction through PCA. To curb overfitting on uneven and heterogeneous data, we place Principal Component Analysis (PCA) inside the preprocessing stage, which lowers the number of features and the computational load while holding on to the information that separates the classes.
– Locality-aware attention transferred to radiography. We adopt the MLLKSA block of Luong et al., developed for breast histopathology, and fine-tune it on chest X-rays (CXRs). The contribution here is the transfer and its empirical assessment, not the mechanism itself.
– A transformer and CNN ensemble. To bring together their differing inductive biases, we join the fine-tuned Vision Transformer with convolutional networks in an ensemble setup. The transformer branch handles long-range global structure while the convolutional branch is stronger on local detail, and together they lift robustness and generalization.
– The ability to explain everything in detail. In order to guarantee that model decisions are clear, clinically reliable, and robust to shortcut learning, we incorporate a multi-method interpretability pipeline that uses gradient-based and model-agnostic visualization tools.

Design, Simulation, and Sizing Assessment of an Off-Grid Solar PV–PEM Green Hydrogen Generation and Compressed Storage System for Bangladesh

Existing PV–hydrogen studies often emphasize component modeling or techno-economic assessment, while integrated Bangladesh-specific frameworks remain limited. This study develops a MATLAB/Simulink-based solar PV–PEM system
using NASA POWER solar data, integrating PV generation, MPPT-controlled DC–DC conversion, PEM electrolysis, and compressed hydrogen storage. Mathematical verification is also performed for hydrogen production and storage performance.
Although LCOH is beyond the present scope, the results indicate strong potential for solar-driven hydrogen production in Bangladesh.

Archives: Paper Submission