Align, Generate, Learn: A Novel Closed-Loop Framework for Cross-Lingual In-Context Learning

TL;DR

This paper introduces a self-supervised, closed-loop framework for cross-lingual in-context learning that improves multilingual task performance without external retrievers or fine-tuning, demonstrating state-of-the-art results.

Contribution

It presents a novel internal example selection method using a retrieval-generation alignment and semantic coherence losses, enhancing scalability and generalizability.

Findings

Achieves state-of-the-art performance on multilingual benchmarks.

Demonstrates robustness across diverse language families.

Outputs are fluent, relevant, and semantically correct.

Abstract

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.