Automatic Short Answer Grading via Multiway Attention Networks

TL;DR

This paper introduces a multiway attention network for automatic short answer grading that effectively models semantic relations between student and reference answers, improving accuracy in diverse K-12 domains.

Contribution

It presents a novel end-to-end framework utilizing multiway attention to better understand and score free-text student answers across multiple domains.

Findings

Outperforms state-of-the-art baselines on real-world K-12 dataset

Effectively captures semantic relations between answers

Demonstrates robustness across diverse open-ended questions

Abstract

Automatic short answer grading (ASAG), which autonomously score student answers according to reference answers, provides a cost-effective and consistent approach to teaching professionals and can reduce their monotonous and tedious grading workloads. However, ASAG is a very challenging task due to two reasons: (1) student answers are made up of free text which requires a deep semantic understanding; and (2) the questions are usually open-ended and across many domains in K-12 scenarios. In this paper, we propose a generalized end-to-end ASAG learning framework which aims to (1) autonomously extract linguistic information from both student and reference answers; and (2) accurately model the semantic relations between free-text student and reference answers in open-ended domain. The proposed ASAG model is evaluated on a large real-world K-12 dataset and can outperform the state-of-the-art baselines in terms of various evaluation metrics.