Universal Information Extraction with Meta-Pretrained Self-Retrieval
This addresses the problem of extracting diverse and complex information structures in NLP for researchers and practitioners, offering a novel retrieval-based approach that is not incremental.
The paper tackles the challenge of Universal Information Extraction (Universal IE) by proposing MetaRetriever, which retrieves task-specific knowledge from pretrained language models to enhance extraction of complex structures, achieving new state-of-the-art results on 4 IE tasks and 12 datasets across various scenarios.
Universal Information Extraction~(Universal IE) aims to solve different extraction tasks in a uniform text-to-structure generation manner. Such a generation procedure tends to struggle when there exist complex information structures to be extracted. Retrieving knowledge from external knowledge bases may help models to overcome this problem but it is impossible to construct a knowledge base suitable for various IE tasks. Inspired by the fact that large amount of knowledge are stored in the pretrained language models~(PLM) and can be retrieved explicitly, in this paper, we propose MetaRetriever to retrieve task-specific knowledge from PLMs to enhance universal IE. As different IE tasks need different knowledge, we further propose a Meta-Pretraining Algorithm which allows MetaRetriever to quicktly achieve maximum task-specific retrieval performance when fine-tuning on downstream IE tasks. Experimental results show that MetaRetriever achieves the new state-of-the-art on 4 IE tasks, 12 datasets under fully-supervised, low-resource and few-shot scenarios.