Bridging the Linguistic Divide: A Survey on Leveraging Large Language Models for Machine Translation

2025-04-02

AI Summary

This paper addresses key bottlenecks in low-resource machine translation (MT)—including scarce parallel corpora, limited linguistic tools, and constrained computational resources—by systematically surveying the paradigm evolution of large language models (LLMs) for MT. Methodologically, it integrates few-shot prompting, cross-lingual transfer, parameter-efficient fine-tuning (e.g., LoRA), LLM-driven synthetic data generation (e.g., back-translation, lexical augmentation), and native quality evaluation metrics. It is the first to comprehensively identify LLM-MT–specific challenges: hallucination, bias propagation, and evaluation inconsistency—and empirically delineates performance boundaries between LLM-based and traditional encoder-decoder MT across multilingual settings. The work contributes a principled, inclusive, and scalable MT development framework; provides a reproducible technical decision guide with benchmark comparisons; and establishes foundations for robust, trustworthy MT in low-resource scenarios. (149 words)

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

The advent of Large Language Models (LLMs) has significantly reshaped the landscape of machine translation (MT), particularly for low-resource languages and domains that lack sufficient parallel corpora, linguistic tools, and computational infrastructure. This survey presents a comprehensive overview of recent progress in leveraging LLMs for MT. We analyze techniques such as few-shot prompting, cross-lingual transfer, and parameter-efficient fine-tuning that enable effective adaptation to under-resourced settings. The paper also explores synthetic data generation strategies using LLMs, including back-translation and lexical augmentation. Additionally, we compare LLM-based translation with traditional encoder-decoder models across diverse language pairs, highlighting the strengths and limitations of each. We discuss persistent challenges such as hallucinations, evaluation inconsistencies, and inherited biases while also evaluating emerging LLM-driven metrics for translation quality. This survey offers practical insights and outlines future directions for building robust, inclusive, and scalable MT systems in the era of large-scale generative models.

Problem

• Leveraging LLMs for low-resource machine translation
• Addressing data scarcity with synthetic generation techniques
• Evaluating LLM strengths/limitations vs traditional MT models

Innovation

• Few-shot prompting for low-resource adaptation
• Synthetic data generation via back-translation
• Parameter-efficient fine-tuning techniques