Distillation for Multilingual Information Retrieval

TL;DR

This paper introduces Multilingual Translate-Distill (MTD), an extension of the Translate-Distill framework, to improve multilingual information retrieval by training models that effectively rank documents across multiple languages.

Contribution

The work extends Translate-Distill to support multilingual document collections, demonstrating significant performance improvements over previous training methods in MLIR tasks.

Findings

MTD outperforms previous state-of-the-art by 5-25% in nDCG@20

MTD achieves 15-45% improvements in MAP

Model robustness to language mixing in training batches

Abstract

Recent work in cross-language information retrieval (CLIR), where queries and documents are in different languages, has shown the benefit of the Translate-Distill framework that trains a cross-language neural dual-encoder model using translation and distillation. However, Translate-Distill only supports a single document language. Multilingual information retrieval (MLIR), which ranks a multilingual document collection, is harder to train than CLIR because the model must assign comparable relevance scores to documents in different languages. This work extends Translate-Distill and propose Multilingual Translate-Distill (MTD) for MLIR. We show that ColBERT-X models trained with MTD outperform their counterparts trained ith Multilingual Translate-Train, which is the previous state-of-the-art training approach, by 5% to 25% in nDCG@20 and 15% to 45% in MAP. We also show that the model is robust to the way languages are mixed in training batches. Our implementation is available on GitHub.