Abstract
This paper presents a new corpus and annotation guideline for a novel coreference resolution task on fictional texts, and analyzes its unique characteristics. FantasyCoref contains 211 stories of Grimms’ Fairy Tales and 3 other fantasy literature annotated in the omniscient writer’s point of view (OWV) to handle distinctive aspects in this genre. This task is more challenging than general coreference resolution in two ways. First, documents in our corpus are 2.5 times longer than the ones in OntoNotes, raising a new layer of difficulty in resolving long-distant referents. Second, annotation of literary styles and concepts raise several issues which are not sufficiently addressed in the existing annotation guidelines. Hence, considerations on such issues and the concept of OWV are necessary to achieve high inter-annotator agreement (IAA) in coreference resolution of fictional texts. We carefully conduct annotation tasks in four stages to ensure the quality of our annotation. As a result, a high IAA score of 87% is achieved using the standard coreference evaluation metric. Finally, state-of-the-art coreference resolution approaches are evaluated on our corpus. After training with our annotated dataset, there was a 2.59% and 3.06% improvement over the model trained on the OntoNotes dataset. Also, we observe that the portion of errors specific to fictional texts declines after the training.
Highlights
This may be due to the nature of this entertainment genre being not as important
According to the published code of EE5, the number of candidate spans is limited to OntoNotes model, only the result of the best model 5https://github.com/lxucs/coref-ee
300 for the reason that the implementation of entity equalization (EE) requires O(k2) memory with k being the number of mentions extracted, while other higher-order inference (HOI) approaches requires O(k)
Summary
A group of three linguists is formed for this project, who use an open source tool called CorefAnnotator (Reiter, 2018) to create our FantasyCoref corpus. Our annotation guidelines are largely based on the OntoNotes Coreference Guidelines 7.0 (Hovy et al, 2006), while referring to other studies on literary texts (Bamman et al, 2020; Roesiger et al, 2018) to consider the characteristics of the genre. While defining coreference relations, we adopt the entity-cluster view, which is adopted by the CoNLL shared tasks (Pradhan et al, 2011, 2012). This view indicates that two or more mentions are non-hierarchically grouped into the same entity cluster (Cranenburgh, van, 2019). To the best of our knowledge, this is the first publicly available corpus that provides full coreference annotation on fantasy literature. We believe that this work will lead a new perspective of coreference resolution on this unexplored domain.
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call