Ellipsis Resolution as Question Answering: An Evaluation
This work addresses ellipsis resolution in natural language processing, providing a novel method that significantly outperforms existing approaches, though it is incremental in applying QA techniques to a specific linguistic task.
The paper tackled ellipsis resolution by framing it as a question answering problem, achieving state-of-the-art results with F1 scores improving from 70.00 to 86.01 for Sluice Ellipsis and from 72.89 to 78.66 for Verb Phrase Ellipsis.
Most, if not all forms of ellipsis (e.g., so does Mary) are similar to reading comprehension questions (what does Mary do), in that in order to resolve them, we need to identify an appropriate text span in the preceding discourse. Following this observation, we present an alternative approach for English ellipsis resolution relying on architectures developed for question answering (QA). We present both single-task models, and joint models trained on auxiliary QA and coreference resolution datasets, clearly outperforming the current state of the art for Sluice Ellipsis (from 70.00 to 86.01 F1) and Verb Phrase Ellipsis (from 72.89 to 78.66 F1).