ReLiK: Retrieve and LinK, Fast and Accurate Entity Linking and Relation Extraction on an Academic Budget

Entity Linking (EL) and Relation Extraction (RE) are fundamental tasks inNatural Language Processing, serving as critical components in a wide range ofapplications. In this paper, we propose ReLiK, a Retriever-Reader architecturefor both EL and RE, where, given an input text, the Retriever module undertakesthe identification of candidate entities or relations that could potentiallyappear within the text. Subsequently, the Reader module is tasked to discernthe pertinent retrieved entities or relations and establish their alignmentwith the corresponding textual spans. Notably, we put forward an innovativeinput representation that incorporates the candidate entities or relationsalongside the text, making it possible to link entities or extract relations ina single forward pass and to fully leverage pre-trained language modelscontextualization capabilities, in contrast with previousRetriever-Reader-based methods, which require a forward pass for eachcandidate. Our formulation of EL and RE achieves state-of-the-art performancein both in-domain and out-of-domain benchmarks while using academic budgettraining and with up to 40x inference speed compared to competitors. Finally,we show how our architecture can be used seamlessly for Information Extraction(cIE), i.e. EL + RE, and setting a new state of the art by employing a sharedReader that simultaneously extracts entities and relations.