EvaAI: A Multi-agent Framework Leveraging Large Language Models for Enhanced Automated Grading

In recent times, Massive Open Online Courses (MOOCs) have become increasingly popular for offering accessible and versatile learning opportunities to a broad audience. However, specific tasks that are necessary in such systems and usually require human involvement, like grading assignments, can be difficult to automate and scale. Furthermore, recent studies have highlighted the capabilities of Large Language Models (LLMs) in a variety of natural language processing tasks, yet the effectiveness of these models in evaluating assignments in educational settings like MOOCs is still an under-researched field. In this paper, we introduce a novel multi-agent architecture, powered by LLMs and the AutoGen framework, that aims in automating the grading of subject-agnostic student assignments. Additionally, we present interaction examples between the agents during grading coding assignments, shedding light on the system's potential to mimic human-like grading nuances and incorporate the tutor’s feedback. This research aims to demonstrate the advancements in automated grading, emphasizing the role of multi-agent systems in educational technology.