To address the increasing demand for personalized instruction in education and ameliorate the time-consuming and inflexible nature of traditional lesson planning, this paper presents the development and evaluation of an intelligent, AI-based lesson preparation tool. The system integrates Large Language Models (LLMs), including the Langchain-Chatchat framework, ChatGLM3, and ERNIE-3.5-8K, with the objective of significantly enhancing both the efficiency of teacher preparation and the quality of instruction. The tool implements four core functionalities: rapid question-answering based on a local knowledge base, one-click intelligent generation of instructional images and PowerPoint presentations (PPTs), and automated assignment generation and grading. Results from functional testing and user feedback evaluations indicate that the tool operates stably, effectively reduces educators' preparation time, enriches the dimensions of teaching content, and increases student engagement. The successful implementation of this research provides a valid paradigm for the deep application of artificial intelligence technology in the education sector, demonstrating its substantial potential in advancing personalized learning and automating pedagogical tasks.

In response to the urgent demand for high-calibre, multidisciplinary software engineering professionals arising from the development of new engineering disciplines, and to address critical bottlenecks in traditional software engineering education—such as the disconnect between theory and practice and insufficient student agency—this research endeavours to establish a systematic, quantifiable blended teaching quality enhancement strategy. Guided by the logic of Outcome-Based Education (OBE), this paper undertakes a comprehensive and systematic backward redesign of the software engineering programme's talent development objectives, curriculum design, teaching implementation, and assessment framework. The core innovation lies in proposing and implementing a tripartite, deeply integrated teaching model centred on Project-Based Learning (PBL), encompassing online, offline, and practical components, alongside a fully integrated quality monitoring and continuous feedback mechanism. Empirical application and data analysis demonstrate that this approach significantly enhances the actual attainment of students' Course Learning Outcomes (CLOs). Particularly in core engineering competency metrics—engineering practice, complex problem-solving, and team collaboration—student performance shows substantial and systematic improvement, with teaching quality assessment indicators effectively enhanced. Research confirms that the blended teaching model guided by OBE principles represents an effective, actionable, and quantifiable pathway for enhancing the quality of software engineering talent cultivation. It provides crucial theoretical underpinnings and practical reference value for curriculum reform and implementation within comparable engineering education domains.

This paper takes the undergraduate graduation design of the Computer Science and Technology major as the research object, and fully records the entire process of the project "Design and Implementation of a Rapid Goods Recognition System for Cabinets Based on Deep Learning" from initiation to implementation. It focuses on analyzing the key actions and core gains in the stages of topic selection and decision-making, technical preparation, engineering practice, and reflection and summary. In the topic selection stage, it breaks through the limitations of traditional development directions and finally selects the uncontacted deep learning field among small program development, conventional system design, and deep learning applications. In the technical preparation stage, it builds a theoretical framework for target detection and system development based on pre-graduate study during the winter vacation and literature review. In the engineering practice stage, aiming at the problem of insufficient computing power of personal equipment, it systematically explores server rental and selection schemes, and overcomes a series of technical difficulties in environment setup, code reproduction, and function integration. Finally, a cabinet goods recognition system with image detection, video recognition, and result export functions is completed. The research confirms that this graduation design not only realizes the cross-field integration of professional knowledge but also cultivates the abilities of independent learning, problem diagnosis, and engineering implementation. It provides a reusable graduation design practice paradigm for computer major students and helps them efficiently complete the core tasks in the final stage of their studies.

To address the theory-practice gap in ethics instruction within artificial intelligence (AI) general education, this action research study developed a three-layer pedagogical model—"Problem Orientation, Implementation, and Value Shaping"—enhanced by AI technology. Centered on the core case study, "My Professional Assistant," and supported by complementary cases, the model was implemented over one semester. A mixed-methods approach, comprising pre- and post-test questionnaires, content analysis of student deliverables, and in-depth interviews, was employed for evaluation. Results demonstrated significant improvements in students' sensitivity to technological ethics (p < 0.001) and their knowledge of responsible AI practices (p < 0.001). Qualitative findings revealed that students transitioned from being mere tool users to responsible supervisors, integrating ethical considerations deeply into their professional practice. Furthermore, several student projects were adopted by external partners, generating impact beyond the classroom. This research provides an actionable framework and empirical support for the systematic cultivation of professional ethics literacy in general AI education.

Aiming at the problems existing in the traditional teaching of "Principles and Applications of Database Systems" for software engineering majors in private application-oriented undergraduate universities—such as overemphasis on theory while neglecting practice, significant differences in students' foundational knowledge, and a single evaluation method—this study integrated the CDIO (Conceive-Design-Implement-Operate) engineering education model with the OBE (Outcome-Based Education) concept, and implemented a systematic teaching reform using the "Student Performance Management System" as the core teaching case. Following the CDIO process of "Conceive-Design-Implement-Operate" and relying on SQL Server as the database tool, student-centered project-based teaching was carried out. Adhering to the logic of "reverse design and forward implementation", a three-dimensional objective system covering knowledge, competence, and quality was reconstructed. Progressive project-based teaching was adopted, along with a differentiated strategy of "stratified grouping + flexible tasks", and a multi-dimensional assessment system consisting of "process evaluation (60%) + summative evaluation (40%)" was established. Teaching practice shows that this model effectively stimulates students' learning initiative, and significantly enhances their engineering practice and teamwork abilities. The CDIO-OBE integrated model can effectively improve the teaching quality of application-oriented courses in private universities, providing a referable path for the reform of similar courses.

Gamma-ray burst (GRB) detection is crucial for triggering rapid follow-up observations and enabling subsequent multi-wavelength studies, yet traditional methods are limited by their difficulty in modeling long-range temporal dependencies in noisy, non-stationary data. In this work, we propose GRBNet, a self-attention–based neural network for GRB detection from same-source multi-detector time-tagged event (TTE) sequences. GRBNet leverages multi-head self-attention to adaptively aggregate discriminative evidence over the full observation window, explicitly capturing long-term dependencies and multi-pulse structures, and integrates complementary responses from multiple detectors under a unified tokenization scheme to improve robustness against background drift and low SNR in individual instruments. Experiments on real Fermi/GBM observations show strong detection performance (recall=1.00, precision=1.00) under a stringent triggering threshold 0.99, indicating reliable performance for weak and morphologically complex GRB events.

To address the issues of uneven texture distribution, background mis-migration, and training imbalance in traditional CycleGAN for style transfer of non-paired horse and zebra images, this study proposes an improved model integrating dynamic attention mechanisms, semantic segmentation constraints, and adaptive training strategies. By embedding lightweight space-channel hybrid attention modules in generator residual blocks, the model enhances feature extraction in target regions. A lightweight semantic segmentation network is introduced to enforce local style transfer constraints, preventing redundant background migration. A two-stage adaptive loss weight adjustment strategy is designed to improve training stability. Experiments on the horse2zebra dataset using the PyTorch platform demonstrate that the improved model generates zebra stripes that conform to the subject structure, with non-target region pixel changes below 15.8%. Compared to the original CycleGAN, the model achieves an 8.3% SSIM improvement and 1.5dB PSNR enhancement. This model effectively resolves core limitations of traditional methods, providing a superior solution for style transfer of non-paired animal images.

To address the challenges in efficiently modeling and simulating Electromagnetic Compatibility (EMC) for complex electrical and electronic systems, a system-level EMC analysis method based on cascaded multi-port network theory is proposed. According to the topological structure of the system, this method decomposes the complex system into multiple cascaded multi-port network modules, where electromagnetic energy transmission and coupling occur via ports. In this method, the Electromagnetic Interference (EMI) transfer characteristics of each module are described by the impedance matrix of its internal circuitry. The voltage and current coupling relationships at system ports are calculated using the cascaded impedance matrix of the multi-stage network combined with specific boundary conditions, thereby significantly simplifying the modeling and simulation process. Numerical validation performed on a three-stage cascaded system across the 10–100 MHz frequency band demonstrates that the proposed method achieves high fidelity. Compared with full-wave simulations, the maximum absolute error for the current response is limited to 0.7 mA (occurring at 10 MHz), and for the voltage response is merely 0.037 V, with the average current error maintained at approximately 0.1 mA. The relative error for both responses at critical ports consistently remains within 0.2%. Furthermore, both modeling complexity and calculation time are substantially reduced, providing a feasible technical approach for the EMC modeling and simulation of complex systems.

In response to the limitations of traditional obstacle avoidance algorithms for unmanned vehicles in dealing with unknown obstacles and complex dynamic environments, this paper proposes Q Mixing Network and Video Delivery Network reinforcement learning algorithms, specifically for the research of obstacle avoidance decision-making for unmanned vehicles. By constructing a mapping relationship between local function values and global function values, it is possible to guide obstacle avoidance decisions for unmanned vehicles based on the decomposed function values. Experimental verification was conducted on a simulation platform based on ROS+Gazebo, and compared with the Quantile Regression for Reinforcement Learning algorithm in the same testing environment. The results showed that QMIX and VDN algorithms were more adaptable to complex map environments during training, with obstacle avoidance success rates increased by 16.9% and 18.1%, respectively, effectively improving the obstacle perception and avoidance capabilities of unmanned vehicles.