AICoFe: AI-Based Collaborative Feedback System for Higher Education
AICoFe (AI-based Collaborative Feedback) is a system designed to improve peer feedback quality in higher education by using a human-centered AI approach. It orchestrates a multi-LLM pipeline with GPT-4.1-mini, Gemini 2.5 Flash, and Llama 3.1 to synthesize rubric data and qualitative observations into actionable feedback. A teacher-in-the-loop workflow allows educators to curate AI-generated drafts via Learning Analytics dashboards. The system employs hybrid SQL and MongoDB storage for traceability and semi-structured data management. The paper details its implementation and deployment.
Key facts
- AICoFe stands for AI-based Collaborative Feedback.
- The system uses a multi-LLM pipeline including GPT-4.1-mini, Gemini 2.5 Flash, and Llama 3.1.
- It synthesizes quantitative rubric data and qualitative observations.
- A teacher-in-the-loop mediation workflow lets educators curate AI-generated feedback.
- Specialized Learning Analytics dashboards are used by educators.
- Data infrastructure uses hybrid SQL and MongoDB strategy.
- The system aims to bridge inconsistent quality of student-generated comments.
- The paper is published on arXiv with ID 2605.04740.
Entities
Institutions
- arXiv