基于通用大语言模型的计算机系统创新实验设计

[Objective] Disruptive intelligent technologies, such as large language models, are rapidly advancing and reshaping productivity, posing challenges to traditional knowledge-driven teaching models and setting new expectations for talent cultivation in higher education. Integrating these large language models with existing course content and designing new experiments to achieve talent cultivation goals in the intelligent era is crucial. This involves improving students’ digital literacy and training various abilities. [Methods] To address these challenges, this study focuses on talent cultivation ability goals in the context of new productivity. The study analyzes the technical characteristics and applications of current large language models, explores the practical skills that students need, and sets corresponding cultivation goals. This study introduces the concept of “intelligent collaborative innovation ability,” which involves cultivating students’ innovation skills using artificial intelligence technologies such as large language models. This study discusses how to internalize this ability across different majors and proposes an innovative experimental design method characterized by “predetermined direction, diverse paths, and dynamic results.” This method involves setting clear yet flexible goals for real engineering problems and encouraging students to seek solutions from various perspectives using diverse methods. This design tests students’ innovative thinking and strengthens their adaptability to uncertainties and challenges. A dynamic result evaluation mechanism ensures comprehensive and fair assessment, promoting deeper thinking and continuous improvement. [Results] To address these challenges, this study begins with a focus on talent cultivation in the era of new productivity. The study analyzes the technical characteristics and applications of current large language models, explores the practical skills that students need, and sets corresponding cultivation goals. This study introduces the concept of “intelligent collaborative innovation ability,” which involves cultivating students’ innovation skills using artificial intelligence technologies such as large language models. This study discusses how to internalize this ability across different majors and proposes an innovative experimental design method characterized by “predetermined direction, diverse paths, and dynamic results.” This method involves setting clear yet flexible goals for engineering problems and encouraging students to seek solutions from various perspectives using diverse methods. This design tests students’ innovative thinking and strengthens their adaptability to uncertainties and challenges. A dynamic result evaluation mechanism ensures comprehensive and fair assessment, promoting deeper thinking and continuous improvement. [Conclusions] The study can transform teaching methods to focus on students and learning, effectively cultivating innovation and autonomous learning abilities. This method allows for personalized cultivation and a hierarchical evaluation. The proposed ability cultivation direction and experimental design method can serve as a reference model for innovative teaching practices across various disciplines when integrated with large language models.