Utilizing Latent Semantic Word Representations for Automated Essay Scoring

Automated essay scoring (AES) utilizes a set of features to measure the writing quality of essays. However, due to the limits of the existing natural language processing techniques, current AES systems are only capable of making use of shallow text features such as the essay length and the number of the clause. In this paper, we argue that the current AES systems can be further improved by taking into account the latent semantic features. To this end, on top of the commonly used shallow features, we propose three deep semanitc features based on Continuous Bag-of-Words Model (CBOW) and Recursive Auto encoder Model. We use Support Vector Machine for Ranking (SVM^rank) to learn a rating model and test the performance of three new features. Experiments on the publicly available English essay dataset, Automated Student Assessment Prize (ASAP), show that our proposed features are beneficial to automated essay scoring.