Pivot Through English: Reliably Answering Multilingual Questions without Document Retrieval

TL;DR

This paper introduces a method to answer questions in low-resource languages by leveraging strong English question-answering models and a pivoting approach, avoiding the need for document retrieval or language-specific supervision.

Contribution

The paper proposes RM-MIPS, a novel pivot-based retrieval method that outperforms existing baselines in multilingual question answering without target language training data.

Findings

RM-MIPS outperforms baselines by 2.7% on XQuAD and 6.2% on MKQA

Effective in low-resource languages with distractors and distribution shifts

Enables rapid, off-the-shelf multilingual question answering without additional training

Abstract

Existing methods for open-retrieval question answering in lower resource languages (LRLs) lag significantly behind English. They not only suffer from the shortcomings of non-English document retrieval, but are reliant on language-specific supervision for either the task or translation. We formulate a task setup more realistic to available resources, that circumvents document retrieval to reliably transfer knowledge from English to lower resource languages. Assuming a strong English question answering model or database, we compare and analyze methods that pivot through English: to map foreign queries to English and then English answers back to target language answers. Within this task setup we propose Reranked Multilingual Maximal Inner Product Search (RM-MIPS), akin to semantic similarity retrieval over the English training set with reranking, which outperforms the strongest baselines by 2.7% on XQuAD and 6.2% on MKQA. Analysis demonstrates the particular efficacy of this strategy over state-of-the-art alternatives in challenging settings: low-resource languages, with extensive distractor data and query distribution misalignment. Circumventing retrieval, our analysis shows this approach offers rapid answer generation to almost any language off-the-shelf, without the need for any additional training data in the target language.