Expert Systems最新文献

The sixth-generation communications networks (6G) are expected to be deployed in the 2030s with integrated space-aerial-ground and undersea architecture to provide seamless global connectivity. In this architecture, unmanned aerial vehicles (UAVs) are one of the most unique characteristics and are becoming increasingly important. The flexibility, high speed, and infrastructure independence of UAV systems make them ideal for many applications. However, these advantages also create great challenges in effective communication between UAVs. To address these challenges, reinforcement learning (RL) algorithms such as Q-Learning have been investigated. However, the traditional Q-learning algorithm mainly relies on delay parameters in the reward function for decision-making route selection. Aiming to optimise the selection of sustainable and efficient communication for UAVs, we propose an improved routing algorithm based on Q-Learning for UAV communication. Our method integrates latency, energy consumption, and link quality parameters into the reward function to make smarter routing decisions. The simulation results show that Q-Proposed achieves significant gains in terms of packet delivery ratio and end-to-end delay compared to other methods, paving the way for sustainable 6G UAV communications.

Traditional transmission line inspection methods face limitations including low efficiency, high costs, and significant safety risks. Although unmanned aerial vehicle inspection has become a mainstream trend, the efficient processing of the massive amounts of image data it generates, as well as the accurate identification and spatial localization of multi-type and multi-scale defects against complex backgrounds, represents key technical challenges currently faced by the intelligent transmission operation and maintenance. To address these challenges, the study proposes and constructs an end-to-end intelligent safety supervision system for unmanned aerial vehicles. Firstly, this system achieves standardised and automated collection of inspection data through autonomous flight route planning. Secondly, to ensure data security, a localization algorithm that integrates unmanned aerial vehicle flight control data is designed. Experimental results demonstrate that our proposed method achieves outstanding performance. Initially, the accurate retrieval model constructed for massive amounts of unmanned aerial vehicle inspection data in the power grid achieves a retrieval accuracy exceeding 75%. Building on this, the core high-precision defect detection model performs outstandingly, with an average detection rate reaching 83%. Specifically, the detection rate for channel defects is 85%, for unclear text and images on signs (ancillary facilities) is 80%, for damaged lightning rods (ancillary facilities) and damaged protective caps (foundations) is 79%, and for damaged armour rods (hardware) is 72%, verifying the model's effectiveness in identifying multiple types of defects. The research work establishes a complete technological chain from unmanned aerial vehicle data processing and intelligent defect detection to precise spatial localization. The proposed method meets practical application requirements in terms of both identification accuracy and category breadth.

Generative artificial intelligence (GAI) has emerged as a transformative force in STEM education, offering new possibilities for personalised instruction, content creation and interactive learning. This review examines the current landscape of GAI applications in science, technology, engineering and mathematics, highlighting key tools such as GPT-4, DALL$��\cdot �$E, AlphaCode and CodeGen. The paper synthesises recent research and practices to identify the pedagogical benefits of GAI, including enhanced self-paced learning, improved access to resources and support for interdisciplinary instruction. It also addresses critical challenges, such as the reliability of generated content, ethical concerns, data privacy and teacher preparedness. These challenges were identified through a synthesis of recent empirical studies, policy reports and expert commentaries in the field of GAI in education, which consistently highlight these issues as major barriers to effective implementation. Based on these findings, the review describes implications for curriculum integration, professional development and institutional policy. Furthermore, the study situates GAI within a broader historical and theoretical context, tracking its evolution from traditional machine learning and deep learning approaches and aligning its educational applications with constructivist, cognitive load and personalised learning theories. By categorising specific use cases across STEM disciplines, such as automated scientific explanations, AI-generated visualisations, intelligent tutoring systems (ITS), virtual engineering labs and adaptive math assessments, the review illustrates the diverse and practical utility of GAI in classroom and remote learning environments. These cases were selected to represent a cross-section of the core instructional needs in STEM education: explanation, visualisation, guidance, experimentation and assessment, where GAI offers distinct functional advantages. They were categorised according to the primary instructional role they fulfil, allowing a pedagogically meaningful organisation of GAI capabilities aligned with common learning processes in STEM. The analysis also emphasises the importance of teacher agency, student participation and equitable access in shaping effective GAI adoption. Finally, this review identifies key future research directions, including the need for longitudinal studies on learning outcomes, efforts to improve the transparency and explainability of GAI models in educational contexts, the development of domain-specific generative tools tailored to STEM subfields, and the exploration of collaborative human–AI learning environments.

News communication is not only a process of information transmission, but also a means of expressing and shaping ideologies. Journalists often embed different political biases in news reports on the same event, depending on their perspectives. Therefore, detecting political bias in news has become a key tool for understanding media bias and uncovering hidden communication strategies. It helps identify politically biased news at its source, thereby reducing the media's influence on public perception. This paper proposes a tree-structured hierarchical model for detecting political bias in news, where elements such as titles, bodies and sentences serve as nodes. A new method is introduced to extract the central sentence of the news, forming a primary-secondary relationship between sentences. Experimental results show that our model is highly effective in detecting political bias in news.

Artificial intelligence (AI) has transformed numerous sectors by offering innovative solutions to complex problems. However, effective implementation of AI projects requires a systematic and integrative approach to stay current with the latest developments in the field. Despite these improvements, realising successful AI and data science initiatives involves employing systematic approaches that can flexibly respond to changing technology and data environments. Two techniques are used to elucidate the life cycle of a high-level data science project. The six-phase methodology termed CRISP-DM (Cross Industry Standard Process for Data Mining) precisely illustrates the data science life cycle. However, the entire workflow conducted by data scientists is performed using the OSEMN methodology (Obtain-Scrub-Explore-Model-iNterpret). Here, the CRISP-DM and OSEMN frameworks have been evaluated and compared. An empirical study has been conducted experimenting with four study scenarios, each producing enlightening results with respect to the model fit and the prediction rate. According to four case studies, CRISP-DM provides a more accurate and efficient method. All issues taken into account, this study contributes to better understanding the most effective approaches for choosing and applying data mining procedures, giving researchers and practitioners guidance on the most appropriate approach for their data analytic tasks. By comparing and contrasting these approaches, this study adds to the discussion of best practices within the data science community for researchers and practitioners when it comes to selecting the most suitable framework for data analysis.

Accurately predicting the default status of listed companies is crucial for risk management and investment decisions. This study develops a default prediction model for Chinese-listed companies by combining news headlines and numerical indicators. A CNN–BiLSTM architecture is employed to encode news headline data, and a DNN is used to encode numerical indicators, with a fusion weight applied to integrate the two data types. During training, the weights of the CNN–BiLSTM and DNN encoders are frozen, while the hybrid and output layers are fine-tuned. This approach improves prediction accuracy and robustness by combining textual and numerical data through weighted information fusion. Additionally, the incorporation of an attention mechanism enhances the model's interpretability by allowing it to focus on the most important features in the data. The main findings of the research are: (1) The proposed model, which combines both news headlines and numerical indicators, outperforms models that rely only on numerical data or only on news headlines. (2) The proposed model outperforms nine baseline models (including BERT series, Transformers, CatBoost, etc.) in terms of G-mean, AUC and Type II Error. By reducing Type II error, the model helps minimise potential financial losses for institutions. (3) Key numerical indicators, such as ‘Audit opinion type’, ‘Common Stock Profitability Rate’ and ‘Eligibility for Margin Trading and Securities Lending’, are crucial for predicting both default and nondefault samples. In default samples, important news headline terms include ‘Aggressive expansion’, ‘Trustee management’ and ‘Pledged financing’, whereas nondefault samples are characterised by terms such as ‘Wide market potential’, ‘Stabilization and recovery’ and ‘Opening high and rising steadily’.

In recent years, with the rapid advancement of information and communications technology, consumer demands in the beauty and fashion industry have become increasingly diverse and personalised. To address real-world application challenges, makeup transfer models must be capable of adapting to variations in head poses to ensure accurate makeup transfer across different angles and poses. In response to these challenges, we propose a new progressive makeup transfer model based on spatial style features and feature maps, termed progressive makeup transfer based on style features and feature maps (PMT-SM). The PMT-SM model is built upon a vision transformer (ViT) backbone and integrates several key components: feature pyramid network (FPN), makeup feature localiser (MFL), makeup style extractor (MSE) and progressive makeup generator (PMGen). This framework is designed to effectively handle large variations in head poses while ensuring precise and adaptable makeup transfer for real-world applications. According to the experimental results, PMT-SM model achieves scores of 0.94, 0.95 and 7.93 in ArcFace similarity, structural similarity index (SSIM) and Fréchet inception distance (FID), respectively, demonstrating superior ability to preserve identity and background information compared with existing makeup transfer models. In addition, the images generated by the proposed model achieve a best-selected ratio (BSR) of 25.00%, indicating that the PMT-SM model provides significant advantages in terms of visual quality and naturalness over other makeup transfer methods in the user study.