Summarization as Indirect Supervision for Relation Extraction

Lu, Kening; Hsu, I-Hung; Zhou, Wenxuan; Ma, Mingyu Derek; Chen, Muhao

doi:10.48550/arxiv.2205.09837

Cited by 3 publications

(6 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…OpenIE Datasets: Given how data-hungry deep learning models are and how costly it is to manually label OpenIE datasets, most OpenIE training sets are weakly labeled using high-confidence extractions from prior OpenIE models to get "silverstandard" labels. For example, the CopyAttention (Cui et al, 2018), SpanOIE (Zhan and Zhao, 2020), and OIE4 (Kolluru et al, 2020b) (Lu et al, 2022).…”

Section: Related Workmentioning

confidence: 99%

“…For backtranslation we use Facebook-FAIR's WMT'19 German-English and English-German models (Ng et al, 2019) and retain only those back translated sentences whose entailment confidence is above 80%. For relation extraction, we use a pretrained SuRE model, a state-of-the-art relation extraction model (Lu et al, 2022) without any additional fine-tuning and keep all relations with confidence above 80%. These confidence thresholds are hyperparameters that may be adjusted.…”

Section: Datasets and Metricsmentioning

confidence: 99%

“…An example of data augmentation via relation extraction. The method used for relation extraction is SuRE(Lu et al, 2022). The sentence is from the OIE4 training set.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Abstractive Open Information Extraction

Pei,

Jindal,

Chang

2023

Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

Open Information Extraction (OpenIE) is a traditional NLP task that extracts structured information from unstructured text to be used for other downstream applications.Traditionally, OpenIE focuses on extracting the surface forms of relations as they appear in the raw text, which we term extractive OpenIE. One of the main drawbacks of this approach is that implicit semantic relations (inferred relations) can not be extracted, compromising the performance of downstream applicationsIn this paper, we broaden the scope of OpenIE relations from merely the surface form of relations to include inferred relations, which we term abstractive OpenIE. This new task calls for the development of a new abstractive OpenIE training dataset and a baseline neural model that can extract those inferred relations. We also demonstrate the necessity for a new semantics-based metric for evaluating abstractive OpenIE extractions. Via a case study on Complex QA, we demonstrate the effectiveness of abstractive OpenIE. 1 Sample Sentence Tokyo, officially Tokyo Metropolis, is the capital city of Japan and one of its 47 prefectures. Extractive {Tokyo; is; the capital city of Japan} OpenIE Extractions {Tokyo; is; one of its 47 prefectures} {Tokyo; is; the capital city of Japan} Abstractive {Tokyo; is officially; Tokyo Metropolis} OpenIE Extractions {Tokyo; is; a prefecture} or {Tokyo; is; one of Japan's 47 prefectures}

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Datasets and Metricsmentioning

confidence: 99%

See 1 more Smart Citation

Abstractive Open Information Extraction

Pei,

Jindal,

Chang

2023

Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

show abstract

“…Zero-shot RE methods apply indirect supervision and convert RE tasks to other NLP tasks (Levy et al 2017;Obamuyide and Vlachos 2018;Sainz et al 2021;Lu et al 2022). For the zero-shot setting, human annotators do not label any samples but are usually asked to generate relation paraphrase templates.…”

Section: Zero-shot Relation Extractionmentioning

confidence: 99%

“…Recently, indirectly supervised methods have taken advantage of pretrained models for other NLP tasks to solve RE tasks in the zero-shot setting. For example, RE tasks have been reformulated as question answering problems (Levy et al 2017), as natural language inference (NLI) tasks (Obamuyide and Vlachos 2018;Sainz et al 2021), and as text summarization tasks (Lu et al 2022). Performance in the zero-shot setting, however, still has a significant gap, and highly relies on the quality and diversity of relation paraphrase templates.…”

Section: Introductionmentioning

confidence: 99%

Improving Distantly Supervised Relation Extraction by Natural Language Inference

Zhou,

Qiao,

et al. 2023

AAAI

View full text Add to dashboard Cite

To reduce human annotations for relation extraction (RE) tasks, distantly supervised approaches have been proposed, while struggling with low performance. In this work, we propose a novel DSRE-NLI framework, which considers both distant supervision from existing knowledge bases and indirect supervision from pretrained language models for other tasks. DSRE-NLI energizes an off-the-shelf natural language inference (NLI) engine with a semi-automatic relation verbalization (SARV) mechanism to provide indirect supervision and further consolidates the distant annotations to benefit multi-classification RE models. The NLI-based indirect supervision acquires only one relation verbalization template from humans as a semantically general template for each relationship, and then the template set is enriched by high-quality textual patterns automatically mined from the distantly annotated corpus. With two simple and effective data consolidation strategies, the quality of training data is substantially improved. Extensive experiments demonstrate that the proposed framework significantly improves the SOTA performance (up to 7.73% of F1) on distantly supervised RE benchmark datasets. Our code is available at https://github.com/kangISU/DSRE-NLI.

show abstract

Enhancing Multi-Document Summarization with Cross-Document Graph-based Information Extraction

Zhang,

Elfardy,

Dreyer

et al. 2023

Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics

View full text Add to dashboard Cite

Information extraction (IE) and summarization are closely related, both tasked with presenting a subset of the information contained in a natural language text. However, while IE extracts structural representations, summarization aims to abstract the most salient information into a generated text summary -thus potentially encountering the technical limitations of current text generation methods (e.g., hallucination). To mitigate this risk, this work uses structured IE graphs to enhance the abstractive summarization task. Specifically, we focus on improving Multi-Document Summarization (MDS) performance by using cross-document IE output, incorporating two novel components:(1) the use of auxiliary entity and event recognition systems to focus the summary generation model and; (2) incorporating an alignment loss between IE nodes and their text spans to reduce inconsistencies between the IE graphs and text representations. Operationally, both the IE nodes and corresponding text spans are projected into the same embedding space and pairwise distance is minimized. Experimental results on multiple MDS benchmarks show that summaries generated by our model are more factually consistent with the source documents than baseline models while maintaining the same level of abstractiveness. 1,2

show abstract

Summarization as Indirect Supervision for Relation Extraction

Cited by 3 publications

References 31 publications

Abstractive Open Information Extraction

Abstractive Open Information Extraction

Improving Distantly Supervised Relation Extraction by Natural Language Inference

Enhancing Multi-Document Summarization with Cross-Document Graph-based Information Extraction

Contact Info

Product

Resources

About