SpaDE: Improving Sparse Representations using a Dual Document Encoder for First-stage Retrieval

TL;DR

SpaDE is a novel dual encoder model that enhances sparse document representations for first-stage retrieval, balancing lexical and semantic matching while maintaining efficiency.

Contribution

It introduces a dual encoder approach with co-training to improve sparse representations, addressing vocabulary mismatch without high inference costs.

Findings

SpaDE outperforms existing uni-encoder models on multiple benchmarks.

It effectively balances lexical and semantic matching.

The co-training strategy enhances training efficiency and effectiveness.

Abstract

Sparse document representations have been widely used to retrieve relevant documents via exact lexical matching. Owing to the pre-computed inverted index, it supports fast ad-hoc search but incurs the vocabulary mismatch problem. Although recent neural ranking models using pre-trained language models can address this problem, they usually require expensive query inference costs, implying the trade-off between effectiveness and efficiency. Tackling the trade-off, we propose a novel uni-encoder ranking model, Sparse retriever using a Dual document Encoder (SpaDE), learning document representation via the dual encoder. Each encoder plays a central role in (i) adjusting the importance of terms to improve lexical matching and (ii) expanding additional terms to support semantic matching. Furthermore, our co-training strategy trains the dual encoder effectively and avoids unnecessary intervention in training each other. Experimental results on several benchmarks show that SpaDE outperforms existing uni-encoder ranking models.