Toward Automatic Discourse Parsing of Student Writing Motivated by Neural Interpretation

Abstract

Providing effective automatic essay feedback is necessary for offering writing instruction at a massive scale. In particular, feedback for promoting coherent flow of ideas in essays is critical. In this paper we propose a state-of-the-art method for automated analysis of structure and flow of writing, referred to as Rhetorical Structure Theory (RST) parsing. In so doing, we lay a foundation for a generalizable approach to automated writing feedback related to structure and flow. We address challenges in automated rhetorical analysis when applied to student writing and evaluate our novel RST parser model on both a recent student writing dataset and a standard benchmark RST parsing dataset.

Defining a New NLP Playground

“…One facet of this challenge comes from identifying and analyzing gaps in a student's understanding or learning. For example, apart from simply scoring essays or responses across discrete dimensions such as fluency or sentence structure or by identifying keyspans (Mathias and Bhattacharyya, 2020;Takano and Ichikawa, 2022;Fiacco et al, 2022), one could use LLMs to determine which parts of a freeform submission indicate a gap and associate it with a learning goal provided by the teacher, without using specific (and costly to create) goldlabeled responses, so that the student has actionable feedback and can work on self-improvement. As part of this work, we need to accurately identify which portions of the response are written by the student as opposed to copied from an AI assistant.…”

A Side-by-side Comparison of Transformers for Implicit Discourse Relation Classification

“…Discourse relation classification identifies one such dimension: the coherence relation between clauses or sentences arising from low-level textual cues (Zhao and Webber, 2022;Webber et al, 2019). This makes the task important to several NLP fields, including multi-party dialogue analysis (Li et al, 2022), social media postings analysis (Siskou et al, 2022), and student literary writing analysis (Fiacco et al, 2022). A discourse relation is often marked with explicit connectives such as but, because, and.…”