Automated Discourse Structure Identification in Student Essays for Writing Improvement System
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Automated discourse structure recognition in student essays is crucial in improving the writing technique through recognition and analysis of key elements of argumentative elements. It allows a customized discussion hence aiding students as well as educators in their academic writing enhancement. Current solutions usually depend on rule-based systems or sequence labeling strategies, which find it difficult to deal with long-context dependencies, non-linear structures and generalization to a variety of writing styles. The limitations decrease the precision of recognition of coherent discourse relations. The new framework uses Graph Neural Networks (GNNs) to do discourse modeling in which the essay parts are represented by nodes and the rhetorical relationships between them by edges. This proves to be a more effective semantic net of complex argumentative paths, global context and semantic dependencies, in comparison with traditional models. It is applied to student essays and produces structured discourse graphs which are used to feed on targeted revisionary feedback and coherence analysis. The results have shown better results in discourse role identification, better recognition of argumentation relationships and more practical writing comments on learners.
Fields of Study & Subjects
Natural Language Processing Techniques
Discourse Analysis in Language Studies
Topic Modeling
Publication details
- Journal / source
- International Conference on Intelligent Systems for Communication, IoT and Security (ICISCoIS)
- Pages
- 1–6
- Publication date
- 8 January 2026
- Language
- EN