Replacing context-specific terms with syntactic descriptors and entity categories allows NLP rules for argumentation-component segmentation that are less context dependent¶

Claim · All claims
Evidence · 1 study · theoretical r2 · q2
Currency · newest study 2012 · not yet checked against its sources · not reviewed by a person · overdue for a recheck since 2015: tool and model accuracy changes quickly

Subclaims¶

q2 i? A pre-processing step replacing context-specific terms with syntactic descriptors and nouns with entity categories (e.g., substituting "location" or "city" for "Sydney") lets software compute rules less dependent on context, addressing over-generalization to specific training sets. → Gerry Stahl 2012

Evidence¶

Gerry Stahl 2012¶

Gerry Stahl. (2012). Cognizing mediating: Unpacking the entanglement of artifacts with collective minds. Computer-Supported Collaborative Learning 7:187–191. https://doi.org/10.1007/s11412-012-9148-x

q2 · i? · theoretical · r2

Editorial description of the approach by Mu, Stegmann, Mayfield, Rosé & Fischer for automating analysis of online collaborative discussion into utterances expressing argumentation components such as claims, evidence, and critique. The editorial calls it "a promising approach to automating such analysis."

"To overcome typical over-generalization to specific training sets, the approach tested here replaces the context-specific terms in a corpus with syntactic descriptors and replaces the nouns with entity categories—e.g., substituting "location" or "city" for "Sydney.""

Discussion¶

Related Claims¶

• Essays within each tool group showed consistent homogeneity across named entities, n-grams, and topic ontology
• Retrained BERT relies on high-impact nouns, verbs, and adjectives for correct predictions, but this reliance is muted on the AF dataset — related