Entity Structure Within and Throughout: Modeling Mention Dependencies for Document-Level Relation Extraction

Xu, Benfeng; Wang, Quan; Lyu, Yajuan; Zhu, Yong; Mao, Zhendong

doi:10.1609/aaai.v35i16.17665

Cited by 106 publications

(38 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…BERT was built on the encoder of Transformer architecture [79] and trained with mask and next sentence prediction tasks, thus can "understand" text sentences and provide a strong base model for downstream tasks. In applications that need representation of input sequence such as sentiment analysis [80,81], intent prediction [82,83], POS Tagging [84], named entity recognition [85][86][87][88], event extraction [89][90][91], relation extraction [88,[92][93][94]. BERT encodes the input embedding sequence with selfattention mechanism in a bidirectional manner, which means a vector at any position can "see" all other vectors and represent its meaning based on the global context without step-by-step pro-jection.…”

Section: Applications Of DL Models In Nlpmentioning

confidence: 99%

N/A Modern Approaches in Natural Language Processing

Tho¹

2022

JCSCE

View full text Add to dashboard Cite

Natural Language Processing (NLP) is one of the major branches in the emerging field of Artificial Intelligence (AI). Classical approaches in this area were mostly based on parsing and information extraction techniques, which suffered from great difficulty when dealing with very large textual datasets available in practical applications. This issue can potentially be addressed with the recent advancement of the Deep Learning (DL) techniques, which are naturally assuming very large datasets for training. In fact, NLP research has witnessed a remarkable achievement with the introduction of Word Embedding techniques, which allows a document to be represented meaningfully as a matrix, on which major DL models like CNN or RNN can be deployed effectively to accomplish common NLP tasks. Gradually, NLP scholars keep developing specific models for their areas, notably attention-enhanced BiLSTM, Transformer and BERT. The births of those models have introduced a new wave of modern approaches which frequently report new breaking results and open much novel research directions. The aim of this paper is to give readers a roadmap of those modern approaches in NLP, including their ideas, theories and applications. This would hopefully offer a solid background for further research in this area.

show abstract

Section: Applications Of DL Models In Nlpmentioning

confidence: 99%

N/A Modern Approaches in Natural Language Processing

Tho¹

2022

JCSCE

View full text Add to dashboard Cite

show abstract

“…Similar to our topic, many recent studies have focused on named entity recognition ( [7], [8], [9], [10]), relation extraction ( [11], [12], [13]), and the joint combination of both ( [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24]). Furthermore, [25] as well as [26] have shown that hierarchically using both tasks in a single pipeline can significantly improve the performance thus indicating that information from one task can be exploited by the other.…”

Section: Related Workmentioning

confidence: 99%

KPI-EDGAR: A Novel Dataset and Accompanying Metric for Relation Extraction from Financial Documents

Deußer¹,

Ali²,

Hillebrand³

et al. 2022

Preprint

View full text Add to dashboard Cite

We introduce KPI-EDGAR, a novel dataset for Joint Named Entity Recognition and Relation Extraction building on financial reports uploaded to the Electronic Data Gathering, Analysis, and Retrieval (EDGAR) system, where the main objective is to extract Key Performance Indicators (KPIs) from financial documents and link them to their numerical values and other attributes. We further provide four accompanying baselines for benchmarking potential future research. Additionally, we propose a new way of measuring the success of said extraction process by incorporating a word-level weighting scheme into the conventional F1 score to better model the inherently fuzzy borders of the entity pairs of a relation in this domain.

show abstract

“…Previous generally effective methods for document-level RE are mainly graph-based models and transformerbased models. Graph-based models (Nan et al, 2020;Zeng et al, 2020Zeng et al, , 2021Xu et al, 2021b) gather entity information for relational inference with graph neural networks, and transformer-based methods (Zhou et al, 2021;Xu et al, 2021a;Zhang et al, 2021;Tan et al, 2022a) implicitly capture long-range dependencies. Recently, (Huang et al, 2022;Tan et al, 2022b) found that a large number of positive relations remain unlabeled in document-level RE datasets, especially unpopular relations.…”

Section: Related Workmentioning

confidence: 99%

A Unified Positive-Unlabeled Learning Framework for Document-Level Relation Extraction with Different Levels of Labeling

Wang¹,

Liu²,

Hu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Document-level relation extraction (RE) aimsto identify relations between entities across multiple sentences. Most previous methods focused on document-level RE under full supervision. However, in real-world scenario, it is expensive and difficult to completely label all relations in a document because the number of entity pairs in document-level RE grows quadratically with the number of entities. To solve the common incomplete labeling problem, we propose a unified positiveunlabeled learning framework − shift and squared ranking loss positive-unlabeled (SSR-PU) learning. We use positive-unlabeled (PU) learning on document-level RE for the first time. Considering that labeled data of a dataset may lead to prior shift of unlabeled data, we introduce a PU learning under prior shift of training data. Also, using none-class score as an adaptive threshold, we propose squared ranking loss and prove its Bayesian consistency with multi-label ranking metrics. Extensive experiments demonstrate that our method achieves an improvement of about 14 F1 points relative to the previous baseline with incomplete labeling. In addition, it outperforms previous state-of-the-art results under both fully supervised and extremely unlabeled settings as well. 1

show abstract

Entity Structure Within and Throughout: Modeling Mention Dependencies for Document-Level Relation Extraction

Cited by 106 publications

References 21 publications

N/A Modern Approaches in Natural Language Processing

N/A Modern Approaches in Natural Language Processing

KPI-EDGAR: A Novel Dataset and Accompanying Metric for Relation Extraction from Financial Documents

A Unified Positive-Unlabeled Learning Framework for Document-Level Relation Extraction with Different Levels of Labeling

Contact Info

Product

Resources

About