References & Citations
Computer Science > Computation and Language
Title: PARE: A Simple and Strong Baseline for Monolingual and Multilingual Distantly Supervised Relation Extraction
(Submitted on 14 Oct 2021 (v1), last revised 6 May 2022 (this version, v2))
Abstract: Neural models for distantly supervised relation extraction (DS-RE) encode each sentence in an entity-pair bag separately. These are then aggregated for bag-level relation prediction. Since, at encoding time, these approaches do not allow information to flow from other sentences in the bag, we believe that they do not utilize the available bag data to the fullest. In response, we explore a simple baseline approach (PARE) in which all sentences of a bag are concatenated into a passage of sentences, and encoded jointly using BERT. The contextual embeddings of tokens are aggregated using attention with the candidate relation as query -- this summary of whole passage predicts the candidate relation. We find that our simple baseline solution outperforms existing state-of-the-art DS-RE models in both monolingual and multilingual DS-RE datasets.
Submission history
From: Kartikeya Badola [view email][v1] Thu, 14 Oct 2021 14:45:51 GMT (386kb,D)
[v2] Fri, 6 May 2022 17:11:44 GMT (751kb,D)
Link back to: arXiv, form interface, contact.