Abstract
dc:description.abstractAuthorship models have historically generalized poorly to new domains because of the wide distribution of author-identifying signals across domains. In particular, the effects of topic and genre are highly domain-dependent and impact authorship performance greatly. This paper addresses a gap in resources for cross-genre authorship datasets by introducing CROSSNEWS, a novel cross-genre dataset that includes formal journalistic articles and casual social media posts. CROSSNEWS is the largest authorship dataset of its kind for both verification and attribution, with comprehensive topic and genre annotations supporting both the authorship verification and attribution tasks. Our experiments on CROSSNEWS demonstrates that current models exhibit poor performance in genre transfer scenarios, underscoring the need for adaptable authorship models. We release the CROSSNEWS dataset publicly to encourage future development and evaluation in this field.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Computer Science
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ma, Marcus
- Advisor dc:contributor.advisor
-
- Xu, Wei
- Committee members dc:contributor.committeemember
-
- Ritter, Alan
- Cadigan, John
- Freitag, Dayne
Subjects
dc:subject × 3Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/78527