From First Draft to Final Insight: A Multi-Agent Approach for Feedback Generation
Jie Cao, Chloe Qianhui Zhao, Xian Chen, Shuman Wang, Christian Schunn, Kenneth R. Koedinger, Jionghao Lin
TL;DR
This paper addresses the challenge of generating high-quality, scalable feedback for large student cohorts using large language models. It proposes a multi-agent G-E-RG framework that generates, evaluates, and regenerates feedback across six first-round prompt-method configurations, guided by two educational theories, and assessed with both human and automatic rubrics. Results show that G-E-RG improves evaluation accuracy, component coverage, and key feedback features, with second-round regeneration yielding near-complete component inclusion and more concise feedback, albeit at higher resource cost. The work provides empirical guidance for implementing iterative, learner-centered automated feedback systems and identifies avenues for improvement, such as retrieval accuracy, human-in-the-loop validation, and more varied non-formulaic feedback.
Abstract
Producing large volumes of high-quality, timely feedback poses significant challenges to instructors. To address this issue, automation technologies-particularly Large Language Models (LLMs)-show great potential. However, current LLM-based research still shows room for improvement in terms of feedback quality. Our study proposed a multi-agent approach performing "generation, evaluation, and regeneration" (G-E-RG) to further enhance feedback quality. In the first-generation phase, six methods were adopted, combining three feedback theoretical frameworks and two prompt methods: zero-shot and retrieval-augmented generation with chain-of-thought (RAG_CoT). The results indicated that, compared to first-round feedback, G-E-RG significantly improved final feedback across six methods for most dimensions. Specifically:(1) Evaluation accuracy for six methods increased by 3.36% to 12.98% (p<0.001); (2) The proportion of feedback containing four effective components rose from an average of 27.72% to an average of 98.49% among six methods, sub-dimensions of providing critiques, highlighting strengths, encouraging agency, and cultivating dialogue also showed great enhancement (p<0.001); (3) There was a significant improvement in most of the feature values (p<0.001), although some sub-dimensions (e.g., strengthening the teacher-student relationship) still require further enhancement; (4) The simplicity of feedback was effectively enhanced (p<0.001) for three methods.