Generative Dense Retrieval: Memory Can Be a Burden

TL;DR

This paper introduces Generative Dense Retrieval (GDR), a new retrieval paradigm combining coarse cluster-level matching with fine-grained document retrieval, addressing memory and scalability issues of traditional generative retrieval methods.

Contribution

The paper proposes GDR, a hybrid retrieval framework that improves memory efficiency and scalability by combining cluster-based matching with intra-cluster document retrieval.

Findings

GDR achieves an average of 3.0 R@100 improvement on NQ dataset.

GDR demonstrates better scalability compared to traditional generative retrieval.

The proposed strategies enhance intra-cluster matching accuracy.

Abstract

Generative Retrieval (GR), autoregressively decoding relevant document identifiers given a query, has been shown to perform well under the setting of small-scale corpora. By memorizing the document corpus with model parameters, GR implicitly achieves deep interaction between query and document. However, such a memorizing mechanism faces three drawbacks: (1) Poor memory accuracy for fine-grained features of documents; (2) Memory confusion gets worse as the corpus size increases; (3) Huge memory update costs for new documents. To alleviate these problems, we propose the Generative Dense Retrieval (GDR) paradigm. Specifically, GDR first uses the limited memory volume to achieve inter-cluster matching from query to relevant document clusters. Memorizing-free matching mechanism from Dense Retrieval (DR) is then introduced to conduct fine-grained intra-cluster matching from clusters to relevant documents. The coarse-to-fine process maximizes the advantages of GR's deep interaction and DR's scalability. Besides, we design a cluster identifier constructing strategy to facilitate corpus memory and a cluster-adaptive negative sampling strategy to enhance the intra-cluster mapping ability. Empirical results show that GDR obtains an average of 3.0 R@100 improvement on NQ dataset under multiple settings and has better scalability.