DocRED
PulseAugur coverage of DocRED — every cluster mentioning DocRED across labs, papers, and developer communities, ranked by signal.
-
New framework tackles noise in document-level relation extraction
Researchers have developed a new framework to address noise and structural inconsistencies in document-level relation extraction datasets. This approach quantifies and enforces structural consistency, particularly focus…
-
New framework CONSISTRE improves LLM relation extraction consistency
Researchers have developed CONSISTRE, a new framework designed to improve document-level relation extraction (DocRE) using large language models (LLMs). The framework addresses the issue of LLMs producing inconsistent o…
-
GLiNER-Relex unifies entity and relation extraction in single NLP model
Researchers have introduced GLiNER-Relex, a novel unified framework designed to simultaneously perform named entity recognition and relation extraction. This approach extends the existing GLiNER architecture, utilizing …