RECON: Relation Extraction using Knowledge Graph Context in a Graph Neural Network

Abstract

In this paper, we present a novel method named RECON, that automatically identifies relations in a sentence (sentential relation extraction) and aligns to a knowledge graph (KG). RECON uses a graph neural network to learn representations of both the sentence as well as facts stored in a KG, improving the overall extraction quality. These facts, including entity attributes (label, alias, description, instance-of) and factual triples, have not been collectively used in the state of the art methods. We evaluate the effect of various forms of representing the KG context on the performance of RECON. The empirical evaluation on two standard relation extraction datasets shows that RECON significantly outperforms all state of the art methods on NYT Freebase and Wikidata datasets. RECON reports 87.23 F1 score (Vs 82.29 baseline) on Wikidata dataset whereas on NYT Freebase, reported values are 87.5(P@10) and 74.1(P@30) compared to the previous baseline scores of 81.3(P@10) and 63.1(P@30).

Beyond modeling: NLP Pipeline for efficient environmental policy analysis

“…For the documents in HTML format, either HTML selectors can be used directly in the scraping tool or the BeautifulSoup4 7 library can be integrated within the scraping tool so that the output of the crawling process is a file in txt format. For the transformation of the pure PDF documents into txt files, there are a few python libraries such as pdfMiner 8 , but these tools are not useful to treat scanned images. To tackle this challenge, we recommend AWS Textract 9 , which is a machine learning service that extracts text from scanned documents.…”

“…Other methods for datasets with fewer and more general relation types ( [51], [22]) leverage a pretrained transformer-based model such as RoBERTa [52], [54]. The embeddings created at this point get passed to a classifier that predicts the correct relation between entities [8], [31].…”