Learning to Retrieve Passages without Supervision

TL;DR

This paper introduces Spider, a novel unsupervised pretraining method for dense retrieval in open-domain question answering, which reduces reliance on labeled data and achieves competitive zero-shot performance.

Contribution

The work presents a new pretraining scheme called recurring span retrieval that leverages pseudo examples for contrastive learning without labeled data.

Findings

Spider outperforms pretrained baselines in zero-shot ODQA tasks.

Hybrid retrieval combining Spider and BM25 surpasses individual methods.

Using Spider as initialization enhances supervised training results.

Abstract

Dense retrievers for open-domain question answering (ODQA) have been shown to achieve impressive performance by training on large datasets of question-passage pairs. In this work we ask whether this dependence on labeled data can be reduced via unsupervised pretraining that is geared towards ODQA. We show this is in fact possible, via a novel pretraining scheme designed for retrieval. Our "recurring span retrieval" approach uses recurring spans across passages in a document to create pseudo examples for contrastive learning. Our pretraining scheme directly controls for term overlap across pseudo queries and relevant passages, thus allowing to model both lexical and semantic relations between them. The resulting model, named Spider, performs surprisingly well without any labeled training examples on a wide range of ODQA datasets. Specifically, it significantly outperforms all other pretrained baselines in a zero-shot setting, and is competitive with BM25, a strong sparse baseline. Moreover, a hybrid retriever over Spider and BM25 improves over both, and is often competitive with DPR models, which are trained on tens of thousands of examples. Last, notable gains are observed when using Spider as an initialization for supervised training.