Scholarlink

<p>The rapid advancement of large language models (LLMs) has opened new frontiers in leveraging artificial intelligence (AI) for diagnosing learning progress. However, the practical application of LLMs in educational learning diagnostics is hindered by their inability to proactively collect and interpret learner data. In this paper, we introduce a novel AI-powered feedback diagnostic system based on LLMs that enhances collaborative learning by simulating tutor–student interactions. Our system comprises two AI agents designed to manage collaborative tasks. The <i>tutor agent</i> employs a prompt-based approach to assess learning progress and conduct initial diagnoses. The <i>learner agent</i> utilizes LLMs to parse tutoring guidelines in conjunction with students’ online learning records, generating simulated dialogues between students and tutors to evaluate the system’s learning diagnostic capabilities. Experimental results demonstrate that our system achieves impressive performance in both evaluating learning progress and making differential diagnoses, though we acknowledge limitations related to LLM inconsistency and overconfidence that must be addressed in future iterations. This work represents a significant step toward the seamless integration of AI into tutoring environments, potentially enhancing the accuracy and accessibility of learning diagnostics across diverse educational contexts.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing collaborative learning through AI-powered feedback systems

  • 期刊论文

  • 发表日期：

    2025年11月28日

查看全文
  • Xin Qi,
  • Jianlong Zhou,
  • Yifei Dong,
  • Fang Chen

摘要

The rapid advancement of large language models (LLMs) has opened new frontiers in leveraging artificial intelligence (AI) for diagnosing learning progress. However, the practical application of LLMs in educational learning diagnostics is hindered by their inability to proactively collect and interpret learner data. In this paper, we introduce a novel AI-powered feedback diagnostic system based on LLMs that enhances collaborative learning by simulating tutor–student interactions. Our system comprises two AI agents designed to manage collaborative tasks. The tutor agent employs a prompt-based approach to assess learning progress and conduct initial diagnoses. The learner agent utilizes LLMs to parse tutoring guidelines in conjunction with students’ online learning records, generating simulated dialogues between students and tutors to evaluate the system’s learning diagnostic capabilities. Experimental results demonstrate that our system achieves impressive performance in both evaluating learning progress and making differential diagnoses, though we acknowledge limitations related to LLM inconsistency and overconfidence that must be addressed in future iterations. This work represents a significant step toward the seamless integration of AI into tutoring environments, potentially enhancing the accuracy and accessibility of learning diagnostics across diverse educational contexts.