DocRED: A Large-Scale Document-Level Relation Extraction Dataset

By Yuan Yao, Deming Ye, Peng Li, Xu Han, Yankai Lin, Zhenghao Liu, Zhiyuan Liu, Lixin Huang, Jie Zhou, Maosong Sun. Annual Meeting of the Association for Computational Linguistics, 2019

584 citations recorded by the source corpus.

Multiple entities in a document generally exhibit complex inter-sentence relations, and cannot be well handled by existing relation extraction (RE) methods that typically focus on extracting intra-sentence relations for single entity pairs. In order to accelerate the research on document-level RE, we introduce DocRED, a new dataset constructed from Wikipedia and Wikidata with three features: (1) DocRED annotates both named entities and relations, and is the largest human-annotated dataset for document-level RE from plain text; (2) DocRED requires reading multiple sentences in a document to extract entities and infer their relations by synthesizing all information of the document; (3) along with the human-annotated data, we also offer large-scale distantly supervised data, which enables DocRED to be adopted for both supervised and weakly supervised scenarios. In order to verify the challenges of document-level RE, we implement recent state-of-the-art methods for RE and conduct a thorough evaluation of these methods on DocRED. Empirical results show that DocRED is challenging for existing RE methods, which indicates that document-level RE remains an open problem and requires further efforts. Based on the detailed analysis on the experiments, we discuss multiple promising directions for future research. We make DocRED and the code for our baselines publicly available at https://github.com/thunlp/DocRED.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/2745fc72e1dd53d1c30f17cf05841b163c2f63c9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Zhiyuan Liu
2. Maosong Sun
3. Minlie Huang
4. Xu Han
5. Jie Zhou
6. Yankai Lin
7. Tianyu Gao
8. Peng Li

Citations in the corpus

• MAVEN: A Massive General Domain Event Detection Dataset (2020, methodology): “To ensure the quality of articles, we follow the previous settings (Yao et al., 2019) to use the introductory sections of the Wikipedia articles for annotation.”
• Reasoning with Latent Structure Refinement for Document-Level Relation Extraction (2020, methodology): “Unlike previous work, we focus on document-level relation extraction datasets (Yao et al., 2019; Li et al., 2016a; Wu et al., 2019) from different domains with a large number of relations and entities, which require understanding a document and performing multi-hop reasoning.”
• Document-Level Relation Extraction with Adaptive Thresholding and Localized Context Pooling (2020, methodology): “Experiments on three document-level relation extraction datasets, DocRED (Yao et al. 2019), CDR (Li et al. 2016), and GDA (Wu et al. 2019b), demonstrate that our ATLOP model significantly outperforms the state-of-the-art methods.”
• ERICA: Improving Entity and Relation Understanding for Pre-trained Language Models via Contrastive Learning (2020, methodology): “Document-level RE For document-level RE, we choose DocRED (Yao et al., 2019), which requires reading multiple sentences in a document and synthesizing all the information to identify the relation between two entities.”
• LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks (2024, unknown): “(Ding et al., 2021), the relation classification dataset DocRED (Yao et al., 2019), the event detection dataset MAVEN (Wang et al., 2020), and the sentiment classification dataset GoEmotions (Demszky et al., 2020).”
• Structured information extraction from scientific text with large language models (2024, background): “State-of-the-art transformer-based implementations of pipeline implementations have been shown to perform document level relation extraction relatively well on a variety of general-knowledge corpora 22 and more specialized domains such as chemical-disease relations 23 and gene-disease relations 24 .”
• Coreferential Reasoning Learning for Language Representation (2020, methodology): “Baselines For QUOREF, we compare our Coref-BERT model with three baseline models: (1) QANet (Yu et al., 2018) combines self-attention mechanism with the convolutional neural network, which achieves the best performance to date without pre-training; (2) QANet+BERT adopts BERT representation as an…”
• Improving Text-to-Text Pre-trained Models for the Graph-to-Text Task (2020, methodology): “To address this challenge, we curate lexicalization templates for each property in the WebNLG and DocRED datasets.”
• More Data, More Relations, More Context and More Openness: A Review and Outlook for Relation Extraction (2020, background): “…(Ntro-duction, 2005), SemEval-2010 Task 8 (Hendrickx et al., 2009), KBP37 (Zhang and Wang, 2015) and TACRED (Zhang et al., 2017); and we have NYT-10 (Riedel et al., 2010), FewRel (Han et al., 2018d) and DocRED (Yao et al., 2019) for distant supervision, few-shot and document-level RE respectively.”
• A Comprehensive Survey on Automatic Knowledge Graph Construction (2023, background): “Beyond systems that deal with heterogeneous data, like web pages and table forms, more attention is being paid to effective methods of tackling complex data – for example, jointly unifying multiple acquisition sub-tasks or solutions that harvest knowledge graphs from long-term contexts [22], noisy d”