Education

Chinese Automated Essay Marking System

Tags: #Education #Singapore #LargeLanguageModel #GenerativeAI

Background

In Singapore, grading Chinese essays is a meticulous process that typically requires several hours per batch, depending on the complexity and length of the essays and the grading criteria. Each essay may take approximately 10-15 minutes to grade, ensuring comprehensive feedback on punctuations used, sentence structures and content. Training a Chinese language teacher in Singapore involves a rigorous program, usually spanning about one to four years. This includes foundational training at the National Institute of Education (NIE), coupled with practical teaching experiences. The curriculum involves effective pedagogical skills, deep cultural knowledge, and proficiency in the Chinese language.

Aiming to optimise the grading process, Analytico Asia look deeper into the technical aspects of Chinese essay grading rubrics. Through training a Generative AI model with marking rubrics, the AI unlocks the potential of automating and possibly streamlining the traditional manual evaluation process in the current practices.

Solution

Analytico Asia has developed a specialized model for grading Chinese essays, focusing on punctuation, sentence structure, and content accuracy. By integrating a Large Language Model (LLM), the model significantly reduces essay grading time to approximately 2-3 minutes per essay. This enhancement not only streamlines the assessment process but also facilitates self-assessment opportunities for learners. From a company standpoint, the National Institute of Education could potentially reduce the number of Chinese language teachers required, optimizing resource allocation.

Video Demo