Using Context to Predict the Purpose of Argumentative Writing Revisions

Abstract

While there is increasing interest in automatically recognizing the argumentative structure of a text, recognizing the argumentative purpose of revisions to such texts has been less explored. Furthermore, existing revision classification approaches typically ignore contextual information. We propose two approaches for utilizing contextual information when predicting argumentative revision purposes: developing contextual features for use in the classification paradigm of prior work, and transforming the classification problem to a sequence labeling task. Experimental results using two corpora of student essays demonstrate the utility of contextual information for predicting argumentative revision purposes.

Identifying Semantic Edit Intentions from Revisions in Wikipedia

“…As the first work to model intentions of revisions, our taxonomy distills and extends existing edit type taxonomies. For instance, our intentions of "elaboration" and "verification" are extensions of "evidence" type proposed by (Zhang and Litman, 2016), and a syntactic category of "information deletion" in could be an instance of our "vandalism" or "simplification" depending on the context.…”

“…In terms of revision intentions, Zhang and Litman (2016) incorporated both argumentative writing features and surface changes from Faigley and Witte (1981) and constructed eight categories of revision purposes, such as claims/ideas, warrant/reasoning/backing, rebuttal/reservation, organization, clarify, etc. Tan and Lee (2014) used revisions to understand statement strength in academic writings.…”

“…Our semantic taxonomy of edit intentions builds on prior literature on collaborative writing (Faigley and Witte, 1981;Fitzgerald, 1987), research on document revision analyses (Bronner and Monz, 2012), studies on edit categories (Daxenberger and Gurevych, 2012;Fong and Biuk-Aghai, 2010), and work on purpose/intention classification (Zhang and Litman, 2016). In order to ensure that our taxonomy captured the intentions that Wikipedians would find meaningful, we set up discussions with a group of 12 interested editors on a Wikipedia project talk page, and iteratively refined our taxonomy based on their feedback.…”

Neural End-to-End Learning for Computational Argumentation Mining

“…AM has applications in legal decision making (Palau and Moens, 2009;Moens et al, 2007), document summarization, and the analysis of scientific papers (Kirschner et al, 2015). Its importance for the educational domain has been highlighted by recent work on writing assistance (Zhang and Litman, 2016) and essay scoring (Persing and Ng, 2015;Somasundaran et al, 2016).…”

Towards Modeling Revision Requirements in wikiHow Instructions

“…Revisions in Wikipedia have been leveraged for various NLP tasks, such as spelling error correction (Ehsan and Faili, 2013;Grundkiewicz and Junczys-Dowmunt, 2014;Zesch, 2012), preposition error correction (Cahill et al, 2013), paraphrasing (Max and Wisniewski, 2010), sentence simplification and compression , textual entailment recognition (Zanzotto and Pennacchiotti, 2010) and lexical simplification (Yatskar et al, 2010). Within this framework, a number of studies have analyzed the type of edits that authors made Gurevych, 2013, 2012;Faruqui et al, 2018;Pfeil et al, 2006;Bronner and Monz, 2012;Liu and Ram, 2011) and their intentions (Yang et al, 2017;Zhang and Litman, 2016). These studies built further upon Faigley and Witte (1981) and Jones (2008).…”