Few-Shot Document-Level Relation Extraction (2022-05-04T00:00:00.000000Z)

TL;DR

This work adapts the state-of-the-art sentence-level method MNAV to the document-level and develops it further for improved domain adaptation and finds FSDLRE to be a challenging setting with interesting new characteristics such as the ability to sample NOTA instances from the support set.

Abstract

We present FREDo, a few-shot document-level relation extraction (FSDLRE) benchmark. As opposed to existing benchmarks which are built on sentence-level relation extraction corpora, we argue that document-level corpora provide more realism, particularly regarding none-of-the-above (NOTA) distributions. Therefore, we propose a set of FSDLRE tasks and construct a benchmark based on two existing supervised learning data sets, DocRED and sciERC. We adapt the state-of-the-art sentence-level method MNAV to the document-level and develop it further for improved domain adaptation. We find FSDLRE to be a challenging setting with interesting new characteristics such as the ability to sample NOTA instances from the support set. The data, code, and trained models are available online (https://github.com/nicpopovic/FREDo).

Authors

Michael Färber

5 papers

Nicholas Popovic

2 papers

TL;DR

Abstract

Authors

References28 items

SAIS: Supervising and Augmenting Intermediate Steps for Document-Level Relation Extraction

Document-level Relation Extraction as Semantic Segmentation

Revisiting Few-shot Relation Classification: Evaluation Data and Classification Schemes

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

Document-Level Relation Extraction with Adaptive Thresholding and Localized Context Pooling

Transformers: State-of-the-Art Natural Language Processing

SciREX: A Challenge Dataset for Document-Level Information Extraction

Dialogue-Based Relation Extraction

FewRel 2.0: Towards More Challenging Few-Shot Relation Classification

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

Matching the Blanks: Distributional Similarity for Relation Learning

RENET: A Deep Learning Approach for Extracting Gene-Disease Associations from Literature

Extracting Multiple-Relations in One-Pass with Pre-Trained Transformers

FewRel: A Large-Scale Supervised Few-Shot Relation Classification Dataset with State-of-the-Art Evaluation

Multi-Task Identification of Entities, Relations, and Coreference for Scientific Knowledge Graph Construction

Decoupled Weight Decay Regularization

Position-aware Attention and Supervised Data Improve Slot Filling

Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour

Prototypical Networks for Few-shot Learning

Matching Networks for One Shot Learning

BioCreative V CDR task corpus: a resource for chemical disease relation extraction

Pack Together: Entity and Relation Extraction with Levitated Marker

One-shot to Weakly-Supervised Relation Classification using Language Models

A Novel Document-Level Relation Extraction Method Based on BERT and Entity Information

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Wikidata: a free collaborative knowledgebase

sciERC ID Wikidata ID hyponym-of P279

, Sylvain Gugger , Mariama Drame , Quentin Lhoest , and Alexander M . Rush . 2020 . Transformers : State - ofthe - Art Natural Language Processing

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names