Teaching a New Dog Old Tricks: Resurrecting Multilingual Retrieval Using Zero-Shot Learning
By Sean MacAvaney, Luca Soldaini, Nazli Goharian. 2019
31 citations recorded by the source corpus.
While billions of non-English speaking users rely on search engines every day, the problem of ad-hoc information retrieval is rarely studied for non-English languages. This is primarily due to a lack of data set that are suitable to train ranking algorithms. In this paper, we tackle the lack of data by leveraging pre-trained multilingual language models to transfer a retrieval system trained on English collections to non-English queries and documents. Our model is evaluated in a zero-shot setting, meaning that we use them to predict relevance scores for query-document pairs in languages never seen during training. Our results show that the proposed approach can significantly outperform unsupervised retrieval techniques for Arabic, Chinese Mandarin, and Spanish. We also show that augmenting the English training collection with some examples from the target language can sometimes improve performance.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Peng Shi
- Xinyu Crystina Zhang
- Jimmy J. Lin
- Richard He Bai
- Qun Liu
- Nandan Thakur
- Mehdi Rezagholizadeh
- Ehsan Kamalloo
In Computer Vision and Pattern Recognition
- Jimmy J. Lin
- Peng Shi
- Qun Liu
- Mehdi Rezagholizadeh
- Xiaoguang Li
- Xinyu Crystina Zhang
- Richard He Bai
- Nandan Thakur
In Information Systems
- Xinyu Crystina Zhang
- Peng Shi
- Jimmy J. Lin
- Qun Liu
- Xiaoguang Li
- Richard He Bai
- Nandan Thakur
- Rodrigo Nogueira
Citations in the corpus
- Making a MIRACL: Multilingual Information Retrieval Across a Continuum of Languages (2022, background): “For these classes of retrieval models, there is an emerging thread of research [17, 18, 23, 29] we hope that MIRACL”
- Cross-Lingual Training of Neural Models for Document Ranking (2020, background): “While we are aware of two previous papers along these lines (Shi and Lin, 2019; MacAvaney et al., 2020), this work ex-”