홈 › Papers

Seq2rel: A sequence-to-sequence-based approach for document-level relation extraction

Motivated by the fact that many relations cross the sentence boundary, there has been increasing interest in document-level relation extraction (RE). Document-level RE requires integrating information within and across sentences, capturing complex interactions between mentions of interacting entities. Most document-level RE methods proposed to date are pipeline-based, requiring entities as input. However, previous work has demonstrated that jointly learning to extract entities and relations can improve performance and be more efficient due to shared parameters and training steps. In this paper, we develop a sequence-to-sequence-based approach that can learn the sub-tasks of document-level RE --- entity extraction, coreference resolution and relation extraction --- in an end-to-end fashion. We evaluate our approach on several datasets, in some cases exceeding the performance of existing methods. Finally, we demonstrate that, under our model, the end-to-end approach outperforms a pipeline-based approach. Our code and models will be made publicly available.

Code (0)

등록된 구현이 없습니다.

Tasks

Similar Papers 제목 키워드 기반