Short Answer

Automated short answer grading (ASAG) aims to efficiently and accurately assess student responses to open-ended questions, reducing the workload on educators while providing timely feedback. Current research heavily utilizes large language models (LLMs), often incorporating retrieval-augmented generation or parameter-efficient fine-tuning techniques to improve accuracy and reduce computational costs, sometimes in conjunction with explainable models like neural additive models. This field is significant because it addresses a critical need in education for scalable and effective assessment, impacting both teaching practices and learning analytics.