Automatic Short Answer Grading Using a LSTM Based Approach

Short Answer Grading is an emerging application of Natural Language Processing and text processing. Automated Short Answer Grading (ASAG) is the process of evaluating student-written short responses using computer techniques like Machine learning. The ASAG task has been studied for a long time, but because of the difficulties in the research, it still attracts attention. One of the primary limitations of ASAG is the scarcity of domain-relevant training data. The job of ASAG can be approached using a variety of methods, which can be broadly divided between traditional methods using hand-crafted features and methods based on deep learning. Due to the growing popularity of this field, researchers have been using Deep Learning Approaches to address this challenge over the past five years. This paper explores the methods of creating an LSTM model, to test how close this approach will bring the machine score to that of the Human Score.