Automatic Summarization for Student Reflective Responses

Educational research has demonstrated that asking students to respond to reflection prompts can improve both teaching and learning. However, summarizing student responses to these prompts is an onerous task for humans and poses challenges for existing summarization methods. From the input perspective, there are three challenges. First, there is a lexical variety problem due to the fact that different students tend to use different expressions. Second, there is a length variety problem that student inputs range from single words to multiple sentences. Third, there is a redundancy issue since some content among student responses are not useful. From the output perspective, there are two additional challenges. First, the human summaries consist of a list of important phrases instead of sentences. Second, from an instructor's perspective, the number of students who have a particular problem or are interested in a particular topic is valuable. The goal of this research is to enhance student response summarization at multiple levels of granularity. At the sentence level, we propose a novel summarization algorithm by extending traditional ILP-based framework with a low-rank matrix approximation to address the challenge of lexical variety. At the phrase level, we propose a phrase summarization framework by a combination of phrase extraction, phrase clustering, and phrase ranking. Experimental results show the effectiveness on multiple student response data sets. Also at the phrase level, we propose a quantitative phrase summarization algorithm in order to estimate the number of students who semantically mention the phrases in a summary. We first introduce a new phrase-based highlighting scheme for automatic summarization. It highlights the phrases in the human summaries and also the corresponding semantically-equivalent phrases in student responses. Enabled by the highlighting scheme, we improve the previous phrase-based summarization framework by developing a supervised candidate phrase extraction, learning to estimate the phrase similarities, and experimenting with different clustering algorithms to group phrases into clusters. Experimental results show that our proposed methods not only yield better summarization performance evaluated using ROUGE, but also produce summaries that capture the pressing student needs.