Element Intervention for Open Relation Extraction

Liu, Fangchao; Yan, Lingyong; Lin, Hongyu; Han, Xianpei; Sun, Le

doi:10.18653/v1/2021.acl-long.361

Cited by 8 publications

(9 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One drawback of matching predicate as the relation is that the predicate extracted by OpenIE is commonly a span from current sentence, which may lead models to take the shortcut by directly matching through words co-occurrence. To eliminate this shortcut, we follow several recent works (Agarwal et al, 2021;Liu et al, 2021) to generate paraphrase texts to match the predicate. Specifically, for extracted triplets, we first wrap them with special markers "[H], [R], [T]" correspond to subject, predicate and object.…”

Section: Triplet-paraphrase Constructionmentioning

confidence: 99%

Pre-training to Match for Unified Low-shot Relation Extraction

Liu¹,

Lin²,

Han³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Low-shot relation extraction (RE) aims to recognize novel relations with very few or even no samples, which is critical in real scenario application. Few-shot and zero-shot RE are two representative low-shot RE tasks, which seem to be with similar target but require totally different underlying abilities. In this paper, we propose Multi-Choice Matching Networks to unify low-shot relation extraction. To fill in the gap between zero-shot and few-shot RE, we propose the triplet-paraphrase metatraining, which leverages triplet paraphrase to pre-train zero-shot label matching ability and uses meta-learning paradigm to learn few-shot instance summarizing ability. Experimental results on three different low-shot RE tasks show that the proposed method outperforms strong baselines by a large margin, and achieve the best performance on few-shot RE leaderboard 1 .

show abstract

Section: Triplet-paraphrase Constructionmentioning

confidence: 99%

Pre-training to Match for Unified Low-shot Relation Extraction

Liu¹,

Lin²,

Han³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…As the previous work (Simon et al, 2019;Hu et al, 2020;Liu et al, 2021), we adopt B 3 (Bagga and Baldwin, 1998), V-measure (Rosenberg and Hirschberg, 2007), and Adjusted Rand Index (ARI) (Hubert and Arabie, 1985) to evaluate different methods. Considering that any of the three metrics can measure the clustering performance from different angles, we take the average of B 3 F1, Vmeasure F1 and ARI for comprehensive evaluation.…”

Section: Evaluation Metricsmentioning

confidence: 99%

“…For fair comparison, all model are trained and evaluated on 10 relation types, same as (Simon et al, 2019;Hu et al, 2020;Liu et al, 2021). We implement our model in PyTorch 3 (Paszke et al, 2017) with transformers package 4 (Wolf et al, 2020).…”

Section: Implementation Detailsmentioning

confidence: 99%

“…Hu et al (2020) leverage BERT to extract relational feature and propose a self-supervised framework to learn relation representations from pseudo labels. Because current methods are unstable and easily collapsed (Simon et al, 2019), Liu et al (2021) solve above-mentioned problems from a causal view and propose element intervention to alleviate the spurious correlations in OpenRE models. However, there are still some hard or semi-hard samples wrongly predicted in the representation space due to the spurious correlations from entities and context to the relation type.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning Discriminative Representations for Open Relation Extraction with Instance Ranking and Label Calibration

Wang¹,

Duan²,

Wu³

et al. 2022

Findings of the Association for Computational Linguistics: NAACL 2022

View full text Add to dashboard Cite

Open relation extraction is the task to extract relational facts without pre-defined relation types from open-domain corpora. However, since there are some hard or semi-hard instances sharing similar context and entity information but belonging to different underlying relation, current OpenRE methods always cluster them into the same relation type. In this paper, we propose a novel method based on Instance Ranking and Label Calibration strategies (IRLC) to learn discriminative representations for open relation extraction. Due to lacking the original instance label, we provide three surrogate strategies to generate the positive, hard negative, and semi-hard negative instances for the original instance. Instance ranking aims to refine the relational feature space by pushing the hard and semi-hard negative instances apart from the original instance with different margins and pulling the original instance and its positive instance together. To refine the cluster probability distributions of these instances, we introduce a label calibration strategy to model the constraint relationship between instances. Experimental results on two public datasets demonstrate that our proposed method can significantly outperform the previous state-of-the-art methods 1

show abstract

“…In other words, the Relation Encoder aims to abstract the notion of relation instance, to provide embeddings that are easier to compare. In recent papers [21,71,29,33,65,73,60], relation embeddings are computed using Pretrained Language Models such as BERT [9,59] or RoBERTa [31]. BERT (and RoBERTa) takes as input some tokenized text, and computes for each input token an embedding, which is representative of the token itself and its context of use.…”

Section: Relation Encodermentioning

confidence: 99%

PromptORE - A Novel Approach Towards Fully Unsupervised Relation Extraction

Genest

Portier

Egyed-Zsigmond

et al. 2022

Proceedings of the 31st ACM International Conference on Information &Amp; Knowledge Management

View full text Add to dashboard Cite

Unsupervised Relation Extraction (RE) aims to identify relations between entities in text, without having access to labeled data during training. This setting is particularly relevant for domain specific RE where no annotated dataset is available and for open-domain RE where the types of relations are a priori unknown. Although recent approaches achieve promising results, they heavily depend on hyperparameters whose tuning would most often require labeled data. To mitigate the reliance on hyperparameters, we propose PromptORE, a "Prompt-based Open Relation Extraction" model. We adapt the novel prompt-tuning paradigm to work in an unsupervised setting, and use it to embed sentences expressing a relation. We then cluster these embeddings to discover candidate relations, and we experiment different strategies to automatically estimate an adequate number of clusters. To the best of our knowledge, PromptORE is the first unsupervised RE model that does not need hyperparameter tuning. Results on three general and specific domain datasets show that PromptORE consistently outperforms state-of-the-art models with a relative gain of more than 40% in B 3 , V-measure and ARI. Qualitative analysis also indicates PromptORE's ability to identify semantically coherent clusters that are very close to true relations.

show abstract

Element Intervention for Open Relation Extraction

Cited by 8 publications

References 30 publications

Pre-training to Match for Unified Low-shot Relation Extraction

Pre-training to Match for Unified Low-shot Relation Extraction

Learning Discriminative Representations for Open Relation Extraction with Instance Ranking and Label Calibration

PromptORE - A Novel Approach Towards Fully Unsupervised Relation Extraction

Contact Info

Product

Resources

About