Discourse Graph Guided Document Translation with Large Language Models
This work addresses the problem of maintaining long-range dependencies and coherence in document-level machine translation for users needing high-quality, efficient translations, representing a novel method for a known bottleneck rather than a foundational advancement.
The paper tackled the challenge of preserving discourse coherence in full document translation with large language models by introducing TransGraph, a framework that models inter-chunk relationships using discourse graphs and selectively conditions translation on relevant graph neighborhoods, resulting in consistent improvements in translation quality and terminology consistency across three benchmarks with six languages and lower token overhead.
Adapting large language models to full document translation remains challenging due to the difficulty of capturing long-range dependencies and preserving discourse coherence throughout extended texts. While recent agentic machine translation systems mitigate context window constraints through multi-agent orchestration and persistent memory, they require substantial computational resources and are sensitive to memory retrieval strategies. We introduce TransGraph, a discourse-guided framework that explicitly models inter-chunk relationships through structured discourse graphs and selectively conditions each translation segment on relevant graph neighbourhoods rather than relying on sequential or exhaustive context. Across three document-level MT benchmarks spanning six languages and diverse domains, TransGraph consistently surpasses strong baselines in translation quality and terminology consistency while incurring significantly lower token overhead.