Early Fusion Strategy for Entity-Relationship Retrieval

TL;DR

This paper introduces an IR-centric early fusion approach for entity-relationship retrieval that handles complex multi-relationship queries by combining unstructured document evidence, demonstrated on Wikipedia and ClueWeb data.

Contribution

It presents a novel early fusion strategy for E-R retrieval that is flexible for various entity and relationship types without predefined schemas.

Findings

Promising results on three E-R query collections

Effective retrieval using Wikipedia and ClueWeb data

Extensible representations for complex queries

Abstract

We address the task of entity-relationship (E-R) retrieval, i.e, given a query characterizing types of two or more entities and relationships between them, retrieve the relevant tuples of related entities. Answering E-R queries requires gathering and joining evidence from multiple unstructured documents. In this work, we consider entity and relationships of any type, i.e, characterized by context terms instead of pre-defined types or relationships. We propose a novel IR-centric approach for E-R retrieval, that builds on the basic early fusion design pattern for object retrieval, to provide extensible entity-relationship representations, suitable for complex, multi-relationships queries. We performed experiments with Wikipedia articles as entity representations combined with relationships extracted from ClueWeb-09-B with FACC1 entity linking. We obtained promising results using 3 different query collections comprising 469 E-R queries.