基于多智能体的人机协同解决复杂学习问题实证研究

This study introduces a multi-agent framework based on large language models, exploring avenues for learners to harness human-machine collaboration to tackle complex learning scenarios effectively. Through comparative discourse analysis between single-agent and multi-agent collaborative interactions, this study discovered that learners in a multi-agent environment spontaneously employ multifaceted questioning strategies to efficiently resolve these complex learning problems. This research not only assists educators in reassessing the current limitations of employing large language models for individual dialogues but also provides practical insights for future implementations of large-scale human-machine interactive communities