Natural Logic-guided Autoregressive Multi-hop Document Retrieval for Fact Verification

TL;DR

This paper introduces a natural logic-guided autoregressive multi-hop document retrieval method for fact verification, reducing memory usage and improving stability over existing approaches by dynamically terminating retrieval based on evidence sufficiency.

Contribution

It presents a novel retrieve-and-rerank approach guided by natural logic, which is more memory-efficient and stable than current dense retrieval methods.

Findings

Competitive performance on FEVER, HoVer, FEVEROUS-S datasets

Uses 5 to 10 times less memory than state-of-the-art methods

Improved stability and human interpretability of retrieval decisions

Abstract

A key component of fact verification is thevevidence retrieval, often from multiple documents. Recent approaches use dense representations and condition the retrieval of each document on the previously retrieved ones. The latter step is performed over all the documents in the collection, requiring storing their dense representations in an index, thus incurring a high memory footprint. An alternative paradigm is retrieve-and-rerank, where documents are retrieved using methods such as BM25, their sentences are reranked, and further documents are retrieved conditioned on these sentences, reducing the memory requirements. However, such approaches can be brittle as they rely on heuristics and assume hyperlinks between documents. We propose a novel retrieve-and-rerank method for multi-hop retrieval, that consists of a retriever that jointly scores documents in the knowledge source and sentences from previously retrieved documents using an autoregressive formulation and is guided by a proof system based on natural logic that dynamically terminates the retrieval process if the evidence is deemed sufficient. This method is competitive with current state-of-the-art methods on FEVER, HoVer and FEVEROUS-S, while using $5$ to $10$ times less memory than competing systems. Evaluation on an adversarial dataset indicates improved stability of our approach compared to commonly deployed threshold-based methods. Finally, the proof system helps humans predict model decisions correctly more often than using the evidence alone.