A Short Answer Grading System in Chinese by CNN

Shih-Hung Wu, Chun-Yu Yeh

Abstract: Short answer question is a common type of test in various level of learning. However, it is not common to test students with the short answer questions online, since automatic short answer grading is not easy. In this paper, we report a short answer grading system based on deep learning model CNN. We test the system on two corpus in Chinese. The first one is translated from a public available corpus in English. We also curate a corpus by ourselves. We get promising result on both data set.