基于智能体回复逻辑的教学交互框架研究OA

With agents entering classroom,agent-supported teaching interaction has become an innovative approach for developing students'higher-order thinking skills.However,existing studies mainly focus on the overall effectiveness of agents'intervention in teaching,while insufficiently revealing the mechanisms of how different response logics influence students'higher-order thinking skills.Therefore,this paper constructed the teaching interaction framework based on agent response logic,proposed three types of agent response logics named constructive,heuristic,and evaluative,and conducted a teaching interaction experiment based on agent response logic to investigate the impact of different response logics on students'higher-order thinking skills,and reveal the path of metacognition involved.The results showed that all three types of agent response logics could effectively improve students'higher-order thinking skills,and presented an improvement effect of"heuristic>constructive>evaluative",but their pathways and advantageous dimensions were different.In addition,the multiple regression analysis further showed that metacognition played an important role in the process of agent response logic influencing higher-order thinking skills,while different response logics acted on specific dimensions of metacognition in different ways.The research in this paper provided clear evidence for the differentiated configuration,complementary synergy,and staged integration of agents in teaching interaction,and offered practical implications for the cultivation of students'higher-order thinking skills.