In permissionless blockchains, Proof-of-Stake (PoS) selects validators in proportion to staked assets, securing the ledger by aligning incentives. However, this proportionality amplifies advantages for large validators through rewards, delegation, and liquid staking, leading to stake concentration and threatening decentralization. To address this issue, we introduce the Dynamic Exponential Stake Weighting (DESW) model, grounded in the principle of weighted probability distribution and equilibrium theory. DESW defines an adaptive validator selection rule that rebalances stake weights according to network-wide inequality, measured by the Gini coefficient. By employing weights in the value domain, DESW reduces the influence of oversized validators as centralization increases, without altering rewards, penalties, or protocol flow. Thus, DESW directly addresses the “rich-get-richer” dynamic in PoS and enhances decentralization. We evaluate DESW across three scenarios that are representative of public PoS networks: (i) stable distributions, (ii) large high-turnover systems, and (iii) adversarial stake injections. Relative to baseline PoS, DESW reduces the Gini coefficient from 0.623 to 0.318 (-48.9 × ) and lower Gini from 0.99 to 0.10. These results indicate that DESW strengthens decentralization and fairness while preserving PoS efficiency, offering a practical drop-in improvement for networks.

Course quality is a critical factor shaping learner experience and institutional success. High-quality courses help learners achieve goals and ensure compliance with education standards. Traditional evaluation relies on expert review after course completion, which is slow, costly, and delays improvement. In Massive Open Online Courses (MOOCs), challenges are greater due to sparse and fragmented learning data. We propose LoDiBi (LOQCA, DeepIFSA, BiLSTM), a framework with three integrated modules. LOQCA automatically labels course quality from learner behavior. DeepIFSA imputes missing values using attention, CutMix, and contrastive learning, making it effective in sparse settings. BiLSTM captures temporal learning patterns to enhance prediction accuracy. Combined, these modules enable early prediction of course quality and provide instructional designers with actionable evidence for timely adjustments. Experiments on real-world MOOC datasets show that LoDiBi outperforms existing methods. Data quality was maximum (Completeness and Consistency reached 1). Balanced classification was achieved (MacroF1, Balanced Accuracy greater than 0.9). Strong agreement with ground-truth labels was confirmed (MCC and Kappa greater than 0.9). Predictive performance was also high (Accuracy, Precision, and Recall between 0.93 and 0.94). LoDiBi provides a scalable solution for automated course evaluation, helping institutions make faster, data-driven decisions to improve and adjust course quality.

Massive Open Online Courses (MOOCs) continue to suffer from high dropout rates, raising concerns for both educational quality and institutional efficiency. Predicting learners’ final outcomes is challenging due to three intrinsic data characteristics: extreme sparsity, complex relational structures, and temporal dynamics. We propose GraphGRU-Ed, a two-stage framework that integrates Graph Convolutional Networks (GCNs) with adversarial learning for robust graph-based imputation, and Gated Recurrent Units (GRUs) for sequential prediction. Learners are classified into five performance levels: A (excellent), B (good), C (average), D (low-performing, high dropout risk), and E (dropout). We emphasize label D, which represents learners who remain active but are at imminent risk of dropping out. Early identification of this group is critical for timely interventions to improve course completion. By exploiting graph topology, GraphGRU-Ed imputes missing features while preserving learner relationships, enabling more reliable sequence modeling. Experiments on MOOCCubeX show GraphGRU-Ed achieves F1=0.92 for label D, accuracy=0.95, and AUC=0.99, consistently outperforming traditional baselines (Mean, Median, KNN) and advanced sequential models. These results demonstrate that the framework provides highly reliable predictions (accuracy), excels at correctly identifying at-risk learners (F1 for label D), and maintains strong discriminative ability between dropout-prone and successful learners (AUC). Its strong early-prediction capability enables MOOC platforms to detect at-risk learners and implement effective interventions to reduce dropout rates.

This study addresses the critical need for effective Big Data Quality Management (BDQM) in education, a field where data quality has profound implications but remains underexplored. The work systematically progresses from requirement analysis and standard development to the deployment of tools for monitoring and enhancing data quality in big data workflows. The study's contributions are substantiated through five research questions that explore the impact of data quality on analytics, the establishment of evaluation standards, centralized management strategies, improvement techniques, and education-specific BDQM adaptations. By addressing these questions, the research advances both theoretical and practical frameworks, equipping stakeholders with the tools to enhance the reliability and efficiency of data-driven educational initiatives. Integrating Artificial Intelligence (AI) and distributed computing, this research introduces a novel multi-stage BDQM framework that emphasizes data quality assessment, centralized governance, and AI-enhanced improvement techniques. This work underscores the transformative potential of robust BDQM systems in supporting informed decision- making and achieving sustainable outcomes in educational projects. The survey findings highlight the potential for automated data management within big data architectures, suggesting that data quality frameworks can be significantly enhanced by leveraging AI and distributed computing. Additionally, the survey emphasizes emerging trends in big data quality management, specifically (i) automated data cleaning and cleansing and (ii) data enrichment and augmentation.

Everyone may now learn more conveniently by using e-learning platforms to study online. For these systems to offer an initial personalized learning path (PLP), a component must function as a content recommendation system (RS). This component must also be able to continuously modify the path to accommodate the learner's learning characteristics and the available learning materials in real time. The provision of highly tailored suggestions is still beset by problems like cold start concerns and data sparsity. Recently, there has been a lot of interest in RS development based on knowledge graphs (KG). KGs can leverage the properties of users and items within a unified graph structure, utilizing semantic relationships among entities to address these challenges and offer more relevant recommendations than traditional methods. In this paper, we provide a KG-based learning path recommendation system to aid in English language acquisition by producing a series of lessons intended to successfully lead learners from their present proficiency level to their desired level. We created a domain KG architecture that includes important idea classes and their connections, especially for preparing English certification examinations. Next, to develop an initial PLP recommendation (PLPR) model, we investigated and used graph data mining algorithms (GAs). Lastly, we devised a method to modify the original PLP's lesson sequence to accommodate the learners' learning characteristics following each real-time interval. With the help of our gathered dataset, consistent experimental conditions, and a chosen set of weights, we assessed our solution using standards like accuracy, efficiency, stability, and execution time.

Personalized course recommendations are essential for enhancing learning in Massive Open Online Courses (MOOCs). However, MOOCs data are typically sparse and incomplete, which limits the effectiveness of graph-based recommendation models. To address this challenge, we propose HetGNN-KGAT, a two-stage framework that integrates Heterogeneous Graph Neural Networks (HetGNN) for imputing missing links and attributes with Knowledge Graph Attention Networks (KGAT) for capturing high-order relations. The contributions of this work are threefold. First, we present a graph-based imputation strategy that mitigates sparsity by enriching MOOCs data through HetGNN. Second, we demonstrate that enriching the graph allows KGAT to operate more effectively, highlighting that the novelty lies not in simply combining two models but in strategically enabling KGAT to exploit sparse MOOCs data. Third, we establish a systematic knowledge discovery pipeline that explicitly evaluates input data quality (completeness and consistency) alongside downstream recommendation performance (MAP, NDCG, Precision, Recall), providing a comprehensive view of how data enrichment improves recommendations. Extensive experiments on the MOOCCubeX dataset show that HetGNN-KGAT consistently outperforms strong baselines in both simulated and real-world scenarios. The framework achieves up to 23.55% relative improvement in MAP, 11.37% in F1-score, and 27.70% in NDCG under simulated conditions, with corresponding gains of 7.82%, 3.73%, and 9.24% in real-world settings. These results highlight that the proposed framework not only strengthens the accuracy of course ranking but also enhances the overall relevance and reliability of personalized recommendations. This confirms the value of coupling data-quality enhancement with graph-based recommendation and offers practical insights for advancing adaptive educational technologies.

Scalability remains a key challenge for layer 1 blockchains. ZK-Rollups, leveraging zero-knowledge proofs, offer a promising layer 2 solution by improving throughput and reducing costs while preserving security. However, the performance of ZK-Rollup still poses a major barrier to practical implementation. The proving circuits in popular applications like ERC-20 transactions are highly complex, often containing a large number of constraints, which directly impacts the computation time and resources required to generate zero-knowledge proofs. This study presents an empirical study on the impact of constraint optimization in Circom on the performance of ERC-20 ZK-Rollups using Groth16. Three optimization levels (-O0, -O1, -O2) are evaluated across transaction batches ranging from 4 to 128, with further exploration up to 512 for specific optimization levels to assess scalability. The results show a trade-off: -O2 reduces constraints by up to 73.2% but increases the compilation time by 213.35% at batch size 128, while -O1 offers a more balanced approach suitable for development stages. The findings confirm that the proof generation time is closely tied to constraint count and complexity. Based on these insights, this study introduces ZCLS (ZK-Circuit Lifecycle Strategy), a practical framework for selecting optimization flags aligned with development stages to enhance ZK-Rollup system efficiency.

In today’s educational landscape, the diverse needs and preferences of learners make Personalized Learning Path Recommendation Systems (PLP-RS) essential for enhancing learning experiences and outcomes. This paper introduces our solution, ODKT-IAC, that integrates Continual Learning (CL) with Adaptive Learning (AL) to ensure the system adapts to continuously changing learner needs. Initially, AL techniques are employed to analyze learning data, including goals, and prior history, to create detailed learning profiles by identifying strengths and weaknesses. Subsequently, CL algorithms, such as online learning in Knowledge Tracing (KT) and incremental Reinforcement Learning (RL), are used to update the model with new learning data, behaviors, and scores. Online learning allows real-time tracking and updating of learners’ knowledge states, while incremental RL enables the algorithm to learn from new data and adjust accordingly, maintaining learner motivation and interest. We evaluate our approach using the Online Learning Activity dataset from Junyi Academy, which includes comprehensive learning logs from 247,548 students over the course of a year. Experimental results demonstrate that our approach enhances learning efficiency in dynamic environments and exhibits superior adaptability and effectiveness compared to traditional methods. Our solution paves the way for intelligent learning recommendation systems (RS), enhancing educational quality, and supporting learners in achieving their academic and career goals.

Blockchain technology has brought a significant advancement in the development of academic certificate management systems by enhancing security, transparency, and decentralization. However, challenges such as certificate revocation, transaction costs, and latency still persist. This research proposes a novel mechanism combining smart contracts and Optimistic Rollups technique to address these issues. By leveraging the off-chain processing feature of Optimistic Rollups, the research has significantly reduced transaction latency and costs in certificate revocation. This integration not only optimizes performance but also maintains transparency and data integrity on the blockchain. Moreover, integrating machine learning for fraud detection not only reinforces the security of the certificate management system but also provides timely alerts before fraudulent transactions occur. The combination of blockchain to ensure decentralization and security, along with machine learning to detect and prevent fraud, creates a comprehensive and advanced certificate management system. The experimental outcomes validate the effectiveness of Optimistic Rollups in certificate revocation, showing a notable approximately 61.92% reduction in both transaction costs and latency. Moreover, the machine learning model displays impressive performance, achieving high accuracy in detecting fraudulent users, with an average F1-score of 99.42% and an AUC score nearing perfection. These results underscore the comprehensive and advanced nature of the certificate management system.

Scholarship administration plays a critical role within higher educational systems, ensuring deserving students have the financial resources to pursue their education. Nevertheless, conventional scholarship administration methods often encounter fraud, mismanagement, restricted accessibility, and the difficult process of authenticating scholarship applicants. The potential of blockchain technology lies in revolutionizing scholarship administration by providing a decentralized, transparent, and secure structure. Despite its potential, blockchain faces scalability challenges, owing to its lower throughput and latency performance compared to non-blockchain systems. This study introduces a decentralized scholarship management model that leverages blockchain technology and zk-rollups, a layer 2 scaling solution. The implementation is on Polygon zkEVM, using Solidity for smart contracts and other contemporary technologies. The experiment conducted by the researchers revealed that the application utilizing zk-rollup technology necessitates transaction costs that are approximately 25 times lower and operates nearly 53

This paper introduces the Distributed Key Architecture (DKA) and elucidates its primary contributions. The proposed architecture combines Shamir's Secret Sharing (SSS) and Distributed Key Generation (DKG) to create a secure and user-friendly blockchain wallet. The advantages of DKA lie in its ability to seamlessly integrate user accessibility with robust security measures. In addition to highlighting the technical aspects, this abstract delves into the contextual application of the methodology, emphasizing its necessity, and outlining the practical benefits it brings. By incorporating Shamir's Secret Sharing and Distributed Key Generation, the DKA not only addresses the current challenges but also redefines the landscape of secure and user-friendly blockchain applications. The experimental results presented herein validate the feasibility of implementing this approach in real-world scenarios, solidifying DKA's potential for practical deployment.

This study aimed to determine the types of sports clubs operating in accordance with the functions and tasks of the Center for Physical Education and Sports at Vietnam National University, Hanoi. Additionally, it proposed solutions to organize sports clubs to enhance the quality of comprehensive physical and mental education and improve sports performance for students. The study employed a mixed-methods design, including document analysis, interviews, social investigations, and pedagogical testing. A total of 70 individuals at Vietnam National University, Hanoi, comprising 40 physical education and sports lecturers and 30 officials and students from the Youth Union and Student Association, were interviewed. Five solutions have been proposed for organizing and operating the university's sports clubs managed by the Center. Additionally, four solutions have been identified to develop sports and entertainment clubs at the member universities of Vietnam National University, Hanoi. These initiatives have led to significant improvements, including enhanced quality of comprehensive physical and mental education, an increase in the number of clubs and members, and notable advancements in sports achievements through training in various disciplines. As a result, the overall sports level at Vietnam National University, Hanoi, has been elevated. The selected solutions have demonstrated effectiveness in promoting sports club activities at the university and advancing physical education as a whole.

The rapid growth of diverse and multi-sourced data has rendered traditional data storage models inadequate to han-dle the sheer volume and complexity. Data Lakes, which store all raw data and all data versions in an easily accessible format, are well-suited for deep data analysis and valuable insights discovery. However, the quality of this data is not guaranteed, raising the question of how to utilize this vast repository effectively. Our research proposes a four-step data quality management process profile, implement, monitor, and improve to oversee and ensure data usability within a data lake. This process employs five commonly used evaluation criteria: accuracy, completeness, consistency, uniqueness, and timeliness. Our study focuses on higher education data, an area that has not been extensively explored in previous research, using real-world data from a uni-versity's computer science department. The application context is managing the quality of input data for a machine-learning model that predicts student graduation outcomes. Two advanced boosting machine learning models, LightGBM and CatBoost, are employed, resulting in a 5% improvement in performance. Our research aims to provide a comprehensive solution for assessing data quality in higher education, saving significant time, effort, and cost while enhancing the reliability of data utilization from data lakes.

Massive Open Online Courses (MOOCs) have become a pioneer in providing access to knowledge for everyone around the world. These courses transcend geographical and linguistic barriers, allowing anyone, anywhere, to learn and enhance their knowledge. Although previous research on recommendation models has shown promising results in course recommendations, building such systems for MOOC platforms still presents significant challenges. Implicit user feedback often lacks explicit negative signals, making it difficult to accurately model user preferences. Additionally, the diversity and sparsity of data, especially for new or niche courses, further hinder traditional methods. This creates a pressing need for new and improved solutions in this field. In this study, we propose H-BERT4Rec, an enhancement of the BERT4Rec model, which leverages Heterogeneous Information Networks (HINs) to address these challenges. HINs integrate diverse data sources and capture complex relationships between entities such as courses, videos, and users. This not only enhances the understanding of user preferences but also strengthens the ability to recommend suitable courses. H-BERT4Rec utilizes Heterogeneous Network Embedding(HNE)-node embedding generation method that leverages HIN to create Pre-train Embeddings, and then improves the BERT4Rec architecture, leading to more accurate and personalized recommendations. We conduct experiments on a real-world MOOC dataset to demonstrate the superior performance of H-BERT4Rec compared to baseline models, achieving an improvement of up to 55,04%. This study contributes a promising new approach for personalized course recommendations in MOOCs, enhancing the learning experience for millions of learners worldwide. This improvement not only promises significant benefits for learners but also opens up new directions for research and development in the field of online education.

In recent years, the widespread adoption of blockchain technology has illustrated its potential for revolutionizing various sectors, particularly within the field of education. In an era when educational institutions confront challenges related to the secure storage, verification, and prevention of academic certificate fraud, blockchain has emerged as a technology well-suited to address these concerns. Besides the problem of storing certificates on the blockchain, revocation has received significant attention from the blockchain community. When incorrect data such as fake certificates or inaccurate personal information stored in a blockchain, revocation becomes an essential function. This article aims to provide technical insights and objective assessments on the performance, security, and privacy aspects of methods for securely revoking academic credentials stored on blockchain platforms. As a result, readers can make informed decisions about which mechanism to implement for their projects or choose to continue further research to optimize these mechanisms.

The rise of online education, particularly through Massive Open Online Courses (MOOCs), has significantly broadened access to high-quality learning resources. However, these platforms still face ongoing challenges in maintaining learner engagement. To address this, researchers have developed advanced personalized Recommendation System (RS), suggesting courses and revealing relationships between learning resources. Given that learners engage with resources in diverse sequences based on their prior knowledge and goals, it is crucial to analyze individual learning behaviors and tailor recommendations accordingly. We propose a Personalized Learning Path (PLP) recommendation method that integrates Knowledge Graph Embedding (KGE) techniques, Collaborative Filtering (CF), and Sequential Pattern Mining (SPM) to recommend video resource sequences. Our approach constructs a MOOC-specific Knowledge Graph (KG), incorporating courses, video resources, and users to analyze learning paths. By vectorizing learners’ chronological patterns, our solution enables detailed representations of various learning trajectories and so can generate highly personalized recommendations. Through a series of experiments with real-world data from XuetangX, we demonstrate the effectiveness of our approach across several key metrics, outperforming state-of-the-art methods. The promising results confirm that our proposed method is highly effective in recommending PLPs and resources in MOOCs.