MuVER: Improving First-Stage Entity Retrieval with Multi-View Entity Representations

TL;DR

MuVER enhances first-stage entity retrieval by creating multi-view representations of entity descriptions, selecting the most relevant view for each mention, leading to state-of-the-art results on key datasets.

Contribution

The paper introduces MuVER, a novel multi-view representation approach that improves entity retrieval by better capturing context-specific mention-entity relations.

Findings

Achieves state-of-the-art performance on ZESHEL.

Improves candidate quality on three Entity Linking datasets.

Demonstrates the effectiveness of multi-view representations.

Abstract

Entity retrieval, which aims at disambiguating mentions to canonical entities from massive KBs, is essential for many tasks in natural language processing. Recent progress in entity retrieval shows that the dual-encoder structure is a powerful and efficient framework to nominate candidates if entities are only identified by descriptions. However, they ignore the property that meanings of entity mentions diverge in different contexts and are related to various portions of descriptions, which are treated equally in previous works. In this work, we propose Multi-View Entity Representations (MuVER), a novel approach for entity retrieval that constructs multi-view representations for entity descriptions and approximates the optimal view for mentions via a heuristic searching method. Our method achieves the state-of-the-art performance on ZESHEL and improves the quality of candidates on three standard Entity Linking datasets