Entity-Duet Neural Ranking: Understanding the Role of Knowledge Graph Semantics in Neural Information Retrieval
This addresses the challenge of enhancing neural ranking models for search systems by incorporating structured knowledge, though it appears incremental as it builds on existing entity-oriented and neural retrieval methods.
The paper tackles the problem of integrating knowledge graph semantics into neural information retrieval by proposing the Entity-Duet Neural Ranking Model (EDRM), which combines entity annotations and interaction-based neural networks, and experiments on a commercial search log show it improves generalization ability.
This paper presents the Entity-Duet Neural Ranking Model (EDRM), which introduces knowledge graphs to neural search systems. EDRM represents queries and documents by their words and entity annotations. The semantics from knowledge graphs are integrated in the distributed representations of their entities, while the ranking is conducted by interaction-based neural ranking networks. The two components are learned end-to-end, making EDRM a natural combination of entity-oriented search and neural information retrieval. Our experiments on a commercial search log demonstrate the effectiveness of EDRM. Our analyses reveal that knowledge graph semantics significantly improve the generalization ability of neural ranking models.