Revisiting Selectional Preferences for Coreference Resolution
This work addresses coreference resolution for natural language processing, but it is incremental as it builds on existing claims about selectional preferences.
The authors tackled the problem of coreference resolution by proposing a dependency-based embedding model for selectional preferences, which improved performance on the CoNLL dataset to match state-of-the-art results, though with associated costs.
Selectional preferences have long been claimed to be essential for coreference resolution. However, they are mainly modeled only implicitly by current coreference resolvers. We propose a dependency-based embedding model of selectional preferences which allows fine-grained compatibility judgments with high coverage. We show that the incorporation of our model improves coreference resolution performance on the CoNLL dataset, matching the state-of-the-art results of a more complex system. However, it comes with a cost that makes it debatable how worthwhile such improvements are.