Domain Adaptation of Coreference Resolution for Radiology Reports

摘要

In this paper we explore the applicability of existing coreference resolution systems to a biomedical genre: radiology reports. Analysis revealed that, due to the idiosyncrasies of the domain, both the formulation of the problem of coreference resolution and its solution need significant domain adaptation work. We reformulated the task and developed an unsuper-vised algorithm based on heuristics for coreference resolution in radiology reports. The algorithm is shown to perform well on a test dataset of 150 manually annotated radiology reports.