Machine Translation (MT) has been predominantly designed for sentence-level translation using transformer-based architectures. While next-token prediction based Large Language Models (LLMs) demonstrate strong capabilities in long-text translation, non-extensive language models often suffer from omissions and semantic inconsistencies when processing paragraphs. Existing preference alignment methods improve sentence-level translation but fail to ensure coherence over extended contexts due to the myopic nature of next-token generation. We introduce Plan2Align, a test-time alignment framework that treats translation as a predictive planning problem, adapting Model Predictive Control to iteratively refine translation outputs. Experiments on WMT24 Discourse-Level Literary Translation show that Plan2Align significantly improves paragraph-level translation, achieving performance surpassing or on par with the existing training-time and test-time alignment methods on LLaMA-3.1 8B.
View on arXiv@article{wang2025_2502.20795,
title={ Plan2Align: Predictive Planning Based Test-Time Preference Alignment in Paragraph-Level Machine Translation },
author={ Kuang-Da Wang and Teng-Ruei Chen and Yu Heng Hung and Shuoyang Ding and Yueh-Hua Wu and Yu-Chiang Frank Wang and Chao-Han Huck Yang and Wen-Chih Peng and Ping-Chun Hsieh },
journal={arXiv preprint arXiv:2502.20795},
year={ 2025 }
}