Efficiency Optimizations for Superblock-based Sparse Retrieval

TL;DR

This paper introduces a simple superblock pruning method for learned sparse retrieval that reduces computational overhead while maintaining retrieval effectiveness, making the process more efficient across various models and datasets.

Contribution

It proposes a novel superblock pruning scheme combined with a compact index, improving efficiency in sparse retrieval without sacrificing relevance.

Findings

Reduces superblock score computation overhead.

Maintains competitive relevance across datasets.

Effective zero-shot configuration for various models.

Abstract

Learned sparse retrieval (LSR) is a popular method for first-stage retrieval because it combines the semantic matching of language models with efficient CPU-friendly algorithms. Previous work aggregates blocks into "superblocks" to quickly skip the visitation of blocks during query processing by using an advanced pruning heuristic. This paper proposes a simple and effective superblock pruning scheme that reduces the overhead of superblock score computation while preserving competitive relevance. It combines this scheme with a compact index structure and a robust zero-shot configuration that is effective across LSR models and multiple datasets. This paper provides an analytical justification and evaluation on the MS MARCO and BEIR datasets, demonstrating that the proposed scheme can be a strong alternative for efficient sparse retrieval.