LAPREL: A Label-Aware Parallel Network for Relation Extraction

Relation extraction is a crucial task in natural language processing (NLP) that aims to extract all relational triples from given sentence. Extracting overlapping complex texts challenging and has received extensive research attention. Most existing methods are based on cascade models employ transform the sentence into vectorized representations. The cascaded structure can cause exposure bias issue; however, representation of each needs be closely related relation with pre-defined types. In this paper, we propose label-aware parallel network (LAPREL) for extraction. To solve issue, apply network, instead framework, table-filling method symmetric pair tagger. obtain task-related embedding, embed prior label information token embedding adjust type. proposed also effectively deal triples. Compared 10 baselines, experiments conducted two public datasets verify performance our network. experimental results show LAPREL outperforms baselines extracting text.