Multi-hop Evidence Retrieval for Cross-document Relation Extraction
About
Relation Extraction (RE) has been extended to cross-document scenarios because many relations are not simply described in a single document. This inevitably brings the challenge of efficient open-space evidence retrieval to support the inference of cross-document relations, along with the challenge of multi-hop reasoning on top of entities and evidence scattered in an open set of documents. To combat these challenges, we propose MR.COD (Multi-hop evidence retrieval for Cross-document relation extraction), which is a multi-hop evidence retrieval method based on evidence path mining and ranking. We explore multiple variants of retrievers to show evidence retrieval is essential in cross-document RE. We also propose a contextual dense retriever for this setting. Experiments on CodRED show that evidence retrieval with MR.COD effectively acquires crossdocument evidence and boosts end-to-end RE performance in both closed and open settings.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Document-level Relation Extraction | CodRED Closed (test) | F1 Score62.53 | 8 | |
| Document-level Relation Extraction | CodRED Closed (dev) | F1 Score0.612 | 8 | |
| Document-level Relation Extraction | CodRED Open (test) | F1 Score57.88 | 7 | |
| Document-level Relation Extraction | CodRED Open (dev) | F1 Score53.06 | 7 |