OneRel: Joint Entity and Relation Extraction with One Module in One Step

Abstract

Joint entity and relation extraction is an essential task in natural language processing and knowledge graph construction. Existing approaches usually decompose the joint extraction task into several basic modules or processing steps to make it easy to conduct. However, such a paradigm ignores the fact that the three elements of a triple are interdependent and indivisible. Therefore, previous joint methods suffer from the problems of cascading errors and redundant information. To address these issues, in this paper, we propose a novel joint entity and relation extraction model, named OneRel, which casts joint extraction as a fine-grained triple classification problem. Specifically, our model consists of a scoring-based classifier and a relation-specific horns tagging strategy. The former evaluates whether a token pair and a relation belong to a factual triple. The latter ensures a simple but effective decoding process. Extensive experimental results on two widely used datasets demonstrate that the proposed method performs better than the state-of-the-art baselines, and delivers consistent performance gain on complex scenarios of various overlapping patterns and multiple triples.

UniRel: Unified Representation and Interaction for Joint Relational Triple Extraction

“…For comparison, We employed twelve strong models as baselines consisting of the SOTA models PRGC (Zheng et al, 2021a), PFN (Yan et al, 2021), TDEER , GRTE (Ren et al, 2021) and OneRel (Shang et al, 2022). We take the experiment results from the original papers of these baselines directly.…”

“…We evaluate the proposed method on two widely used benchmark datasets NYT (Riedel et al, 2010) and WebNLG (Gardent et al, 2017) (Zeng et al, 2018) 61.0 56.6 58.7 37.7 36.4 37.1 GraphRel (Fu et al, 2019) 63.9 60.0 61.9 44.7 41.1 42.9 OrderCopyRE (Zeng et al, 2019) 77.9 67.2 72.1 63.3 59.9 61.6 CasRel BERT (Wei et al, 2020) 89.7 89.5 89.6 93.4 90.1 91.8 TPlinker BERT 91.3 92.5 91.9 91.7 92.0 91.9 PRGC BERT (Zheng et al, 2021a) 93.3 91.9 92.6 94.0 92.1 93.0 R-BPtrNet BERT 93.0 92.1 92.5 93.8 92.4 93.1 GRTE BERT (Ren et al, 2021) 92.9 93.1 93.0 93.7 94.2 93.9 EmRel (Xu et al, 2022) 91.7 92.5 92.1 92.7 93.0 92.9 OneRel BERT (Shang et al, 2022) 92 is produced by distant supervision from New York Times articles which has 24 predefined relations. WebNLG was first created for natural language generation task and is adapted to relational triple extraction by Zeng et al (2018), which has 171 predefined relations.…”

“…To handle the problem, researchers (Zeng et al, 2018(Zeng et al, , 2019Nayak and Ng, 2020;Ye et al, 2021) propose generative models that view triple as a token sequence. Some works Ren et al, 2021;Shang et al, 2022) introduce methods to extract triples in one-stage but suffer from huge prediction space. Other researchers (Wei et al, 2020;Zheng et al, 2021b;Wu and Shi, 2021) decompose RTE into different subtasks, but learn the interaction between sub-tasks only by input sharing, or falling into the cascade error.…”

“…Zheng et al (2021a) decompose RTE into three subtasks. and Shang et al (2022) extract relational triples in one stage to eliminate exposure bias. In general, these works accomplish end-to-end extraction by various ways to factorize and re-assemble the label space of relational triples.…”

“…We can easily determine that London and UK are correlated in the prerequisite of given the interactions of (London-is_capital_of) and (UK-is_capital_of). However, existing works either enumerate all entity-relation-entity triples Shang et al, 2022), which suffer from huge prediction space, or model them with separate modules (Wei et al, 2020;Zheng et al, 2021a; respectively for named entity recognition (NER) and relation extraction (RE), resulting in neglect of interdependencies. In general, the absence of unified modeling of interactions limits these methods from fully utilizing their interdependencies for better extraction.…”