Learning to Translate for Multilingual Question Answering

TL;DR

This paper introduces a learn-to-translate approach for multilingual question answering that optimally combines different translation methods and directions, outperforming traditional translation-based methods on a multilingual forum dataset.

Contribution

It proposes a novel model that learns optimal feature weights for multiple translation methods and directions in multilingual QA, improving over standard translation strategies.

Findings

Learn-to-translate approach outperforms baseline (p<0.05)

Effective in multilingual forum dataset with English, Arabic, Chinese

Combines multiple translation methods and directions for better QA performance

Abstract

In multilingual question answering, either the question needs to be translated into the document language, or vice versa. In addition to direction, there are multiple methods to perform the translation, four of which we explore in this paper: word-based, 10-best, context-based, and grammar-based. We build a feature for each combination of translation direction and method, and train a model that learns optimal feature weights. On a large forum dataset consisting of posts in English, Arabic, and Chinese, our novel learn-to-translate approach was more effective than a strong baseline (p<0.05): translating all text into English, then training a classifier based only on English (original or translated) text.