Online Language Adaptive Sampling for Better Distributed Cross-lingual Gains

arXiv:2609.14969 · cs.CL, cs.AI · Submitted 2026-09-14 · Read on arXiv

cs.CL, cs.AI

Submitted: 2026-09-14

Updated: 2026-09-14

Comments: Accepted to ENMLP 2026 Findings

Code: https://github.com/felixgaschi/multilingual-alignment-and-transfer

License: http://creativecommons.org/licenses/by/4.0/

The gist: Realignment is a promising approach for improving the cross-lingual transfer ability of multilingual language models, particularly for extremely low-resource languages (LRLs).

Terminology

Abstract

Realignment is a promising approach for improving the cross-lingual transfer ability of multilingual language models, particularly for extremely low-resource languages (LRLs). However, existing realignment methods rely on uniform and random sampling of parallel sentences across languages, which may be suboptimal under limited batch sizes. In practice, models may benefit from seeing certain languages more frequently, especially those that are poorly aligned, and the optimal distribution can evolve throughout training. In this work, we propose a simple yet effective adaptive sampling strategy that assigns trainable sampling probabilities to each language. Languages that contribute more to the realignment loss are sampled more frequently in subsequent batches, and the optimal distribution can evolve throughout training. Our method employs an inner-outer optimization loop with a small overhead, leading to consistent performance improvements and, more importantly, distributing the gains across languages. We observed a +0.67 average performance increase on all tasks with XLM-R, and +0.60 with Gemma 2 9B compared with uniform realignment. Furthermore, our method is robust across different models. Code available at https://github.com/felixgaschi/multilingual-alignment-and-transfer.

Sources

Related papers