ReLiK: Retrieve and LinK, Fast and Accurate Entity Linking and Relation Extraction on an Academic Budget
Riccardo Orlando, Pere-Lluis Huguet Cabot, Edoardo Barba, Roberto Navigli

TL;DR
ReLiK introduces a fast, accurate Retriever-Reader architecture for entity linking and relation extraction that leverages pre-trained models efficiently, achieving state-of-the-art results with significantly reduced inference time.
Contribution
The paper presents a novel input representation and a unified architecture for EL and RE, enabling single-pass processing and improved performance on benchmarks.
Findings
Achieves state-of-the-art in-domain and out-of-domain performance.
Up to 40x faster inference compared to existing methods.
Unified approach for entity linking and relation extraction with shared components.
Abstract
Entity Linking (EL) and Relation Extraction (RE) are fundamental tasks in Natural Language Processing, serving as critical components in a wide range of applications. In this paper, we propose ReLiK, a Retriever-Reader architecture for both EL and RE, where, given an input text, the Retriever module undertakes the identification of candidate entities or relations that could potentially appear within the text. Subsequently, the Reader module is tasked to discern the pertinent retrieved entities or relations and establish their alignment with the corresponding textual spans. Notably, we put forward an innovative input representation that incorporates the candidate entities or relations alongside the text, making it possible to link entities or extract relations in a single forward pass and to fully leverage pre-trained language models contextualization capabilities, in contrast with…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
- 🤗sapienzanlp/relik-retriever-e5-base-v2-aida-blink-encodermodel· 590 dl590 dl
- 🤗sapienzanlp/relik-retriever-e5-base-v2-aida-blink-wikipedia-indexmodel· 165 dl· ♡ 1165 dl♡ 1
- 🤗sapienzanlp/relik-retriever-e5-base-v2-blink-first1M-encodermodel· 5 dl5 dl
- 🤗sapienzanlp/relik-reader-deberta-v3-large-aidamodel· 377 dl· ♡ 1377 dl♡ 1
- 🤗sapienzanlp/relik-retriever-small-nyt-question-encodermodel· 265 dl265 dl
- 🤗sapienzanlp/relik-retriever-small-nyt-document-indexmodel· 123 dl123 dl
- 🤗sapienzanlp/relik-reader-deberta-v3-large-nytmodel· 242 dl242 dl
- 🤗sapienzanlp/relik-entity-linking-largemodel· 99 dl· ♡ 1199 dl♡ 11
- 🤗sapienzanlp/relik-reader-deberta-v3-base-aidamodel· 409 dl409 dl
- 🤗sapienzanlp/relik-entity-linking-basemodel· 96 dl· ♡ 596 dl♡ 5
Videos
Taxonomy
TopicsData Quality and Management · Topic Modeling
MethodsSPEED: Separable Pyramidal Pooling EncodEr-Decoder for Real-Time Monocular Depth Estimation on Low-Resource Settings
