Choose the Final Translation from NMT and LLM hypotheses Using MBR Decoding: HW-TSC's Submission to the WMT24 General MT Shared Task

TL;DR

This paper describes HW-TSC's submission to WMT24, combining NMT and LLM hypotheses with MBR decoding for improved English-Chinese translation, achieving competitive results.

Contribution

It introduces the integration of continue pre-training, supervised fine-tuning, and contrastive preference optimization for LLM-based MT, alongside traditional NMT training strategies.

Findings

Achieved competitive translation results in WMT24 evaluation.

Demonstrated effectiveness of MBR decoding in combining NMT and LLM hypotheses.

Enhanced translation quality through novel training techniques for LLMs.

Abstract

This paper presents the submission of Huawei Translate Services Center (HW-TSC) to the WMT24 general machine translation (MT) shared task, where we participate in the English to Chinese (en2zh) language pair. Similar to previous years' work, we use training strategies such as regularized dropout, bidirectional training, data diversification, forward translation, back translation, alternated training, curriculum learning, and transductive ensemble learning to train the neural machine translation (NMT) model based on the deep Transformer-big architecture. The difference is that we also use continue pre-training, supervised fine-tuning, and contrastive preference optimization to train the large language model (LLM) based MT model. By using Minimum Bayesian risk (MBR) decoding to select the final translation from multiple hypotheses for NMT and LLM-based MT models, our submission receives competitive results in the final evaluation.