Align, Generate, Learn: A Novel Closed-Loop Framework for Cross-Lingual In-Context Learning
This provides a scalable and generalizable solution for multilingual tasks, especially benefiting low-resource languages, though it is incremental as it builds on existing in-context learning paradigms.
The paper tackles the problem of cross-lingual in-context learning by proposing a self-supervised framework that uses LLMs to internally select task-relevant examples, achieving state-of-the-art performance on multilingual benchmarks and outperforming existing baselines.
Cross-lingual in-context learning (XICL) has emerged as a transformative paradigm for leveraging large language models (LLMs) to tackle multilingual tasks, especially for low-resource languages. However, existing approaches often rely on external retrievers or task-specific fine-tuning, limiting their scalability and generalizability. In this paper, we propose a novel self-supervised framework that harnesses the generative capabilities of LLMs to internally select and utilize task-relevant examples. Our method introduces two key objectives: a retrieval-generation alignment loss to optimize the quality of selected examples and a semantic coherence loss to ensure cross-lingual consistency. Through extensive experiments on multilingual benchmarks, our approach achieves state-of-the-art performance, significantly outperforming existing baselines. Further analysis highlights its robustness across diverse language families and its ability to generalize to unseen tasks. Human evaluations confirm the superior fluency, relevance, and semantic correctness of outputs generated by our method. This work provides a scalable, effective, and generalizable solution for cross-lingual in-context learning.