Journal of ICT Standardization

Research on the Impact of Social Media Short Videos on College English Listening and Speaking Teaching

Tingting Wu — 2026-06-05

This study focuses on the integration path and effect of short videos on social media and college English listening and speaking teaching, aiming to provide a reference for the reform of college English listening and speaking teaching through empirical exploration. The research adopted a mixed research method (quantitative +
qualitative), selecting 90 second-year students of other majors from a certain university as the research subjects. All students had achieved a level 4 English proficiency (425–500 points) and had not received short video-assisted listening and speaking teaching. They were randomly allocated to the experimental group (45 people). A 16-week teaching experiment was carried out by using short videos from platforms such as YouTube and TikTok for auxiliary teaching and a control group (45 people, taught using the traditional textbook “New Horizons College English (Third Edition)” listening and speaking course). During the experiment, the teaching of the two groups was carried out around three units: campus life, cross-cultural communication, and career planning. The teacher was the same person to control the variables. Data were collected through several methods, including tests, questionnaires on learning motivation, classroom observations, and interviews. After conducting independent sample t-tests and ANOVA with SPSS 24.0 and qualitative coding analysis with Nvivo 12, it was found that: in terms of listening and speaking abilities, the average score of the experimental group after the test was (78.6 ±
7.2), which showed a significant advantage compared with the control group (68.5 ±
8.1) (t=5.82, p<0.01). The improvement rate (16.3 points) was 2.4 times that of the control group (6.7 points). Meanwhile, the difference in the long dialogue question type was the most significant (difference proportion +22.95%). After the test, the average score of the experimental group was (76.3 ±7.8), which showed a significant advantage compared with the control group (66.2 ±8.4) (t=4.95, p<0.01). Compared with the control group, the pronunciation dimension (78.5 ±6.9) was 12.3 points higher and the fluency dimension (77.2 ±7.1) was 11.5 points higher. In terms of learning motivation, compared with the control group (p<0.05), the four ARCS model dimensions of attention (4.2 ±0.5), relevance (4.1 ±0.6), confidence (4.0 ±0.5), and satisfaction (3.9 ±0.6) in the experimental group all had obvious advantages. Among them, the score difference in the confidence dimension is the most significant, at 0.9 points. Moreover, the motivation showed a phased characteristic of rapid increase from the 1st to the 4th week and stable improvement after the 8th week, forming a positive cycle from ability improvement to motivation enhancement, despite the absence of a prominent fluctuation in the motivation of the control group throughout the process (the difference in each dimension was ≤0.1). In terms of teaching adaptability, the real context (multiple accents, life-like scenarios) of short videos solves the problem of disconnection between teaching materials and language resources. The duration of 15 seconds to 5 minutes is suitable for fragmented learning, and the high interactivity (imitation, recording, and mutual evaluation) enhances classroom participation The average number of active speeches per class in the experimental group (18.5 times) was three times that of the control group (6.2 times), and the participation in group discussions (92%) was significantly higher than that of the control group (65%). Further analysis of the effects of different types of short videos reveals that the educational category performed best in capturing listening details (4.5 points) and imitating oral pronunciation (4.4 points), the cultural category scored the highest in stimulating learning motivation (4.6 points), and the daily category had a significant advantage in improving oral fluency (4.5 points). Research has confirmed that short videos on social media can effectively assist college students of other majors in English listening and speaking teaching and stimulate their learning motivation. In teaching, complementary advantages should be achieved by adopting strategies such as using educational videos in the basic stage, daily videos in the application stage, and interspersing cultural videos throughout the process.

Optimal Allocation in Enterprise Human Resource Management through Intelligent Scheduling Algorithms

Yun Wu — 2026-06-05

The paper provides a brief introduction to the mathematical model of project scheduling optimization and the genetic algorithm (GA) used to optimize the human resource allocation scheme. The GA was improved by co-evolution and adaptive genetic parameters. Subsequently, simulation experiments were conducted, and the improved algorithm was also compared with the particle swarm optimization (PSO) algorithm and traditional GA. The results demonstrated that the improved GA converged to stability more quickly during the search for the optimal solution, resulting in a more excellent objective function upon convergence. Overall, the scheme optimized by the improved GA exhibited lower human resource costs and a shorter completion cycle.

A Comprehensive Study on Integration of Big Data and AI in Financial Decision-Making

Karim Elalkaoui — 2026-06-05

The rapid proliferation of large-scale financial data, coupled with advancements in Artificial Intelligence (AI), has significantly transformed modern financial decision-making. This paper presents a comprehensive state-of-the-art review of AI-driven approaches supported by Big Data infrastructure in the financial domain. We analyse recent academic contributions (2023–2025) across machine learning(ML), deep learning and hybrid ensemble techniques applied to forecasting, portfolio optimisation, risk assessment and fraud detection. Emerging data architectures such as streaming frameworks and lakehouse platforms are assessed in terms of their ability to support real-time analytics and large-scale model deployment. We highlight the transition towards multimodal and attention-based models that integrate structured and unstructured data sources, and identify key challenges including concept drift, explainability, privacy and robustness. A detailed case study involving a GPU-accelerated hybrid deep learning and ensemble model for BTC–USD price prediction demonstrates practical benefits and current limitations: the hybrid model achieved an RMSE of 2656.69, a MAPE of 2.14%, and an R2 of 0.9626 on the test set. Although the absolute RMSE reflects the inherent volatility of the asset class, the low MAPE (2.14%) and high R2 confirm the model’s predictive efficacy during regime shifts, highlighting the necessity for future integration of macro-scenarios.

Risk Warning and Decision Support Model for Shield Tunneling Construction in Urban Rail Transit

Wenfeng Cao — 2026-06-05

Risks occur due to variable geological conditions and urban constraints. Inadequate risk identification and management in shield tunneling can result in safety hazards, schedule overruns, and increased project costs, necessitating an effective risk assessment and decision-making framework. This research aims to develop a comprehensive risk warning and decision support model to facilitate early risk detection and informed mitigation strategies during shield tunneling construction in urban rail transit projects. The dataset includes geotechnical conditions, tunneling parameters, environmental factors, operational records, and safety monitoring data. Data sources include multi-sensors, and project logs, enabling comprehensive risk analysis and modeling. These multi-sensor inputs include cutterhead torque sensors, thrust force sensors, slurry pressure sensors, vibration and displacement sensors, as well as ground settlement monitoring instruments. The project logs consist of TBM operational logs, geotechnical investigation records, and safety monitoring logs documenting environmental and structural conditions during tunneling. To ensure data quality, pre-processing methods including Interquartile Range (IQR) for outlier detection and mean imputation for missing values are applied. Predictive risk modeling was conducted using Regularized Random Forest (RRF). Risk thresholds were established in accordance with model outputs and relevant safety standards to enable proactive early warning alerts. The decision support module leverages Proximal Policy Optimization (PPO) to recommend adaptive mitigation actions, such as tunneling parameter adjustments and structural reinforcements. Additionally, it facilitates forecasting of potential outcomes, providing dynamic, real-time feedback to the construction management team for rapid operational response. The proposed model demonstrated robust performance in early risk identification and offered actionable recommendations that enhance safety management and operational efficiency within urban shield tunneling projects. The integrated risk warning and decision support framework provides a technically sound and practical tool to improve risk mitigation efficacy, optimize construction decision-making processes, and promote the safe and efficient advancement of urban rail transit shield tunneling construction.

ETSI Activities in Artificial Intelligence and Data

Markus Mueck — 2026-06-05

Beyond classical Information and Communication Technologies (ICT), the European Telecommunications Standards Institute (ETSI) is developing standards in Artificial Intelligence (AI) and Data to support a broader ecosystem of Information Technology (IT) stakeholders. This work is primarily conducted through two Technical Committees: ETSI TC Securing Artificial Intelligence (TC SAI) and ETSI TC Data solutions (TC DATA). TC SAI focuses on ensuring that AI systems are safe, secure, and societally responsible, including developing baseline cybersecurity requirements grounded in structured design principles across the AI lifecycle. TC DATA complements this work by addressing distributed data solutions, including data in transit, data at rest, and data in process, with particular emphasis on supporting the implementation of the EU Data Act through interoperable data models and ontologies.

As a European Standards Organisation (ESO), ETSI contributes to both European policy priorities and globally applicable standards, including Harmonised Standards that support regulatory compliance within the European Union. ETSI provides an open and inclusive environment for industry, academia, and SMEs to participate in the standardisation process, enabling stakeholders to shape the future development and deployment of AI and data-driven technologies while fostering trust, interoperability, and global adoption.

Research on the Architecture and Implementation of Power Archive Information Platform Based on International and Local Standards

Qin Jie — 2026-06-05

In the digital transformation of the power industry, archive management faces the dual-track coordination dilemma of coexisting international standards and local regulations. This study addresses this challenge by proposing a two-tier standard coordination system of “core standards + local extensions,” effectively resolving implementation conflicts among heterogeneous standards through mechanisms such as metadata mapping and compliance checking. Building on this approach, an innovative “five-layer, three-guarantee” hierarchical decoupling platform architecture covering the entire lifecycle of archives has been designed. It clearly delineates five layers: infrastructure, data resources, core services, application functions, and user interaction, complemented by three guarantee mechanisms for standards, security, and operations, ensuring the platform’s openness and scalability. To meet core business needs such as trusted archival verification, intelligent processing, and dynamic association, the research explores the application of key technologies, including blockchain for ensuring archive authenticity, artificial intelligence for intelligent classification and indexing, and digital twins for modeling dynamic mappings between archives and physical entities. At the implementation level, a four-stage model is proposed, along with a performance evaluation system encompassing infrastructure, process management, and outcome benefits, introducing business value-oriented indicators such as “completeness and accuracy of archive searches” and “cost savings from a single-set management system.” Case studies from Guangdong Power Grid demonstrate that this platform significantly improves the digitization rate of archives and the efficiency of cross-system data exchange, offering a reference solution that holds both theoretical value and practical feasibility for building a systematic and standardized archive information platform for power enterprises.