Moinul Zaber

Accurate short-term load forecasting is essential for modern power distribution, yet selecting optimal predictive architectures across diverse feeder profiles remains a key challenge for local grid operators. This paper presents a multi-feeder comparative evaluation of machine learning architectures using hourly load data (January 2024–April 2026) from six distribution feeders within two operational zones under the Bangladesh Rural Electrification Board. The evaluated models include statistical baselines (ARIMA, SARIMA), linear regression, tree ensembles (Random Forest, XGBoost, LightGBM), and recurrent architectures (LSTM, GRU). Feature engineering includes cyclical temporal encodings, weather parameters, and temporal lags. Experimental results demonstrate that tree-based ensembles consistently outperform both statistical baselines and recurrent neural networks across all feeders. Random Forest achieved the highest performance, with a Mean MAPE of 2.35%, and individual feeder errors reaching as low as 0.83% MAPE and 0.02 MW MAE. Meanwhile, LightGBM demonstrated superior operational deployment efficiency, delivering a highly comparable Mean MAPE of 2.56% with the lowest inference time of 0.55 s, and a minimal memory footprint of 1.2 MB. These findings highlight that feature-engineered tree ensembles provide an optimal balance of predictive performance, scalability, and computational efficiency for real-time distribution grid operations across neighboring utility zones.

Research comparing national artificial intelligence (AI) policy scores the state and its institutions, and rarely asks what a policy commits to where a person meets a system: whether someone subject to an automated decision can learn, correct, refuse or appeal it. We contribute Human-Centered AI Policy Instrument (HCAIPI), an instrument for reading policy at that layer, applied to 245 editions from 101 countries against twelve themes from human-computer interaction (HCI) research, recording how firmly each commitment is stated and whether anything would execute it. Of 13,587 relevant passages, 44.7% state a principle with no implementation detail and 11.4% specify a mechanism, a shape invariant to income band, length and volume. What a person is told or consulted about is instrumented at twice the rate of what they can do, and 65% of specified mechanisms are tasks for government. The lexicon recovers 13% to 54% of what a reader accepts; country figures track length.

Centralized identity and access management (IDAM) systems suffer from sin-gle points of failure, lack of authorization transparency, and susceptibility to in-sider threats and privilege abuse. While role-based access control (RBAC) of-fers structured permission management, its enforcement through centralized pol-icy engines introduces auditability gaps unacceptable in modern distributed service delivery environments. This paper presents BDLock, a blockchain-enabled two-tier privacy-aware federated IDAM platform integrating OAuth 2.0, OpenID Connect (OIDC), and Hyperledger Fabric 2.4. The first tier validates JSON Web Tokens (JWT) issued by Keycloak against a Spring Boot resource server, the second tier enforces immutable scope-based RBAC rights on the Hyperledger Fabric ledger, ensuring every access decision is tamper-proof and auditable. Unlike prior approaches, BDLock uniquely bridges OAuth-authenticated off-chain identities to cryptographic on-chain Fabric wallet identities, satisfying all six STRIDE-modelled threats categories across both Web2 and Web3 identity models. Validated with up to 1,800 concurrent users, BDLock achieves a peak throughput of approximately 200 transactions per second using round-robin load balancing. At high concurrency, it outperforms single-peer fallback by up to 25%. Furthermore, it maintains uninterrupted access control during peer failures, eliminating the single point of failure found in all nine compared state-of-the-art systems.

Measuring multidimensional Quality of Life (QoL) increasingly relies on composite indicators to guide sustainable development policy. However, standard linear aggregation frameworks generally assume full compensability, allowing strong performance in one domain to mathematically mask severe deficits in another. For policymakers, this compensatory masking may conceal structural imbalances across sectors, obscuring localized vulnerabilities critical for targeted interventions. To address this diagnostic gap, we introduce the Sustainable Composite Indicator for Policy Evidence (SCIPE), a granular QoL assessment framework that explicitly penalizes horizontal imbalances across SDG-aligned dimensions. Using Bangladesh’s 64 administrative districts as a representative case study, we construct a multi-domain baseline that integrates traditional socio-economic metrics with satellite-derived night-time light data. SCIPE employs Principal Component Analysis (PCA) for data-driven indicator weighting and the Mazziotta-Pareto Index (MPI) to enforce an explicit variance penalty. Sensitivity analysis against the Human Development Index (HDI) confirms convergent validity while demonstrating SCIPE’s sensitivity to inter-sectoral volatility. Empirical findings reveal that penalizing horizontal variance significantly reconfigures district rankings: rewarding balanced developmental trajectories while exposing hidden trade-offs, such as ecological and public health deficits within affluent industrial corridors. Ultimately, this study demonstrates that aggregation choices fundamentally dictate policy interpretation, underscoring the necessity of evaluating balanced progress rather than relying solely on macro-level averages.

Background: Accurate short-term load forecasting (STLF) is essential for modern power distribution, yet selecting an optimal predictive architecture across diverse feeder profiles remains a key challenge for local grid operators. Method: This paper presents a comprehensive multi-feeder comparative evaluation of machine learning architectures using high-resolution hourly load data spanning January 1, 2024 to April 30, 2026 collected across six distribution feeders from two major operational zones under Bangladesh Rural Electrification Board (BREB): Gazipur PBS-1 and Mymensingh PBS-2. The forecasting models used here are classical statistical baselines (ARIMA, SARIMA), Linear Regression, tree-based ensembles (Random Forest, XGBoost, LightGBM), and recurrent deep learning structures (LSTM, GRU). Advanced feature engineering—incorporating cyclical temporal encodings, meteorological parameters, and temporal lags—was applied to capture localized spatial-temporal load dynamics. Results: Experimental results demonstrate that tree-based ensemble models consistently outperform both traditional statistical baselines and recurrent neural networks across all six feeders. Specifically, Random Forest achieved the highest overall predictive performance, attaining a peak average network accuracy of 97.66% (with a lowest feeder MAPE of 0.83% on Sreepur-8 F2). Meanwhile, LightGBM demonstrated superior operational efficiency, achieving a comparable 97.44% average accuracy with training time of only 3.47 seconds and a minimal memory footprint of 1.2 MB. Conclusion: These findings highlight that tree ensembles combined with feature engineering offer an optimal balance of accuracy, scalability, and computational efficiency for real-time distribution grid operation across neighboring utility zones.

Background: Artificial intelligence (AI) integration in mobile health (mHealth) apps offers health care access opportunities in low-resource settings, yet opaque AI recommendations undermine trust and adoption. Existing explainable AI (XAI) frameworks, designed in Western contexts, fail to address the linguistic, cultural, and infrastructural realities of South Asian populations, creating barriers where users cannot understand AI recommendations, clinicians cannot validate outputs, and developers lack implementation guidance. Thus, understanding explainability requirements among educated, digitally literate populations provides foundational insights for future development of inclusive mHealth technologies. Objective: This study aims to (1) investigate stakeholder perceptions of trust and explainability in AI-driven mHealth in Bangladesh; (2) identify demographic predictors of trust; and (3) develop and propose a context-adapted framework benefiting developers, policymakers, clinicians, and end users in resource-constrained settings. Methods: This study used a sequential mixed methods design that combined a quantitative survey (n=137) with a qualitative phase involving 20 stakeholders. This qualitative cohort consisted of developers (n=4), XAI experts (n=6), and clinicians (n=10) who participated through either focus groups or individual interviews. We used statistical analysis to examine demographic predictors and applied thematic analysis to identify explainability needs specific to each stakeholder group. Results: Education level showed a significant effect on trust ( F _{3, 133} =2.81, P =.042). Completed undergraduate students reported lower trust (mean 3.14, SD 1.10) compared with current undergraduates (mean 3.66, SD 0.93), suggesting that undergraduate completion develops critical evaluation skills that may decrease uncritical acceptance of AI systems. Despite recognizing AI’s utility for preliminary guidance, users emphasized the necessity of human validation and expressed concerns about understanding AI’s decision-making logic. Interviews with different stakeholder groups revealed critical gaps. Developers acknowledged minimal explainability implementation in current mHealth apps, while medical professionals unanimously prioritized clinical judgment over automated outputs and advocated for physician-mediated AI systems. Synthesizing findings across all stakeholder groups revealed five core requirements: (1) Human–AI collaboration and clinical validation, (2) Transparent logic paths, (3) Contextual personalization, (4) Cultural and linguistic relevance, and (5) Trust calibration and ethical safeguards. Conclusions: The framework bridges stakeholder misalignments and offers actionable guidance for design, deployment, and policy alignment in resource-constrained environments. By situating explainability within the sociocultural realities of South Asia, this research advances XAI beyond algorithmic transparency toward equity, inclusion, and user empowerment in digital health.

Artificial intelligence is widely framed as a new regulatory frontier. Yet many of its governance challenges market concentration, infrastructure control, interoperability, and equitable access echo long-standing concerns in telecommunications policy. This paper asks three questions: What is the long-run conceptual architecture of telecommunications governance? How is knowledge production within this field distributed globally? And does AI governance represent rupture or path-dependent continuity?Using state-of-the-art natural language processing, semantic embedding, and network analysis on 3,064 articles published between 1975 and 2025, the study reconstructs the latent structure of telecommunications policy scholarship. The findings reveal a historically layered governance regime characterized by cumulative thematic expansion and strong path dependence. At the same time, knowledge production is structurally concentrated in a small group of Global North countries that dominate collaboration networks and agenda setting.Analysis of AI-related research shows that artificial intelligence diffuses across pre-existing policy domains rather than forming an autonomous regulatory field. Moreover, the geography of AI governance mirrors the historical North-dominated structure of telecommunications research. The results suggest that digital governance evolves through institutional layering, and that AI represents an uneven inheritance of telecom policy reproducing both its regulatory logics and its global asymmetries.