AI technology for human computer collaborative intelligent English teaching under DCCM model

Area of Science:

• Artificial Intelligence in Education
• Educational Technology
• Cognitive Science

Background:

• Traditional English teaching struggles to balance personalized tutoring with humanistic care.
• Integrating AI requires innovative models for effective student engagement and support.

Purpose of the Study:

• To propose and validate a human-computer collaborative intelligent teaching model (Dynamic Cognitive Clustering Model - DCCM).
• To enhance English teaching efficiency and reduce teacher workload through AI integration.

Main Methods:

• Developed a DCCM using an improved K-means++ algorithm with time decay and a sliding window for dynamic student cognitive state clustering.
• Conducted four experiments: dataset clustering (TOEFL11), intervention study (subjunctive mood), comparative teaching models, and long-term follow-up (writing instruction).

Main Results:

• The DCCM achieved a high Adjusted Rand Index (ARI) of 0.87 in clustering.
• Demonstrated significant improvements: 78.3% conflict resolution, 52.1% reduction in structural errors, 37% less teacher workload, 34.2% higher student writing scores, and 51.9% decreased error density.
• 43.3% of students advanced at least one CEFR level.

Conclusions:

• The human-computer collaborative model effectively integrates AI feedback with teacher-led instruction.
• This approach enhances teaching efficiency, reduces teacher stress, and optimizes student learning outcomes.
• Further validation in diverse settings is recommended for broader generalizability.