Embedded Files
Work in Progress
- Eye-Tracking Research on Korean EFL Learners' Paraphrasing Strategies
This study investigates how Korean EFL learners allocate visual attention and employ strategic approaches during sentence-level paraphrasing tasks. Using webcam-based eye-tracking technology (RealEye), the research identifies which parts of a source sentence draw learners’ attention and how these patterns differ between high- and low-performing learners. Eye-tracking data are combined with NLP-based scoring metrics—including semantic similarity, lexical change, and syntactic change—to analyze the relationship between gaze behavior and meaning-preserving strategies. The results aim to contribute to the development of cognitively grounded feedback and instructional models in L2 academic writing.
This study investigates how Korean EFL learners allocate visual attention and employ strategic approaches during sentence-level paraphrasing tasks. Using webcam-based eye-tracking technology (RealEye), the research identifies which parts of a source sentence draw learners’ attention and how these patterns differ between high- and low-performing learners. Eye-tracking data are combined with NLP-based scoring metrics—including semantic similarity, lexical change, and syntactic change—to analyze the relationship between gaze behavior and meaning-preserving strategies. The results aim to contribute to the development of cognitively grounded feedback and instructional models in L2 academic writing.
This research is funded by the Learning Science Research Institute at Seoul National University
(2025 Graduate Student Research Support Program)
Building Automatic Scoring Systems for Paraphrasing Assessment
Large Language Models such as a Customized GPT vs. Natural Language Processing Metrics
I am currently working on building an Automatic Paraphrase Scoring (APS) system to effectively integrate the artificial intelligence (AI) into language assessment, especially writing evaluation for second language (L2) learners. I believe establishing this APS would help reduce the cost and burden on grading large volume of essays for teachers and practitioners in the field of English language teaching.
Page updated
Google Sites
Report abuse