State-of-the-Art: Assessing Semantic Similarity in Automated Short-Answer Grading Systems

The use of semantic in Natural Language Processing (NLP) has sparked the interest of academics and businesses in various fields. One such field is Automated Short-answer Grading Systems (ASAGS) for automatically evaluating responses for similarity with the expected answer. ASAGS poses semantic challenges because the responses of a topic are in the responder’s own words. This study is providing an in-depth analysis of work to improve the assessment of semantic similarity between corpora in natural language in the context of ASAGS. Three popular semantic approaches are corpus- based, knowledge-based, and deep learning are used to evaluate against the conventional methods in ASAGS. Finally, the gaps in knowledge are identified and new research areas are proposed