面向多语言机器翻译的语言感知束搜索解码
Language-Informed Beam Search Decoding for Multilingual Machine Translation
AI总结：
针对多语言NMT模型束搜索解码易产生偏离目标翻译的问题，提出与模型无关、无需额外平行数据的语言感知束搜索算法LiBS，有效提升翻译质量并大幅降低偏离目标率。
摘要：
束搜索解码是自回归神经机器翻译（NMT）模型解码的实际标准方法，包括将目标语言指定为输入的多语言NMT模型。然而，多语言NMT模型解码时通常会产生“偏离目标”的翻译——生成的翻译输出并非预期语言。本文首先对一个性能强劲的多语言NMT模型的偏离目标翻译进行错误分析，明确了这类解码结果在束搜索过程中的产生机制。随后我们提出语言感知束搜索（LiBS），这是一种将现成的语言识别（LiD）模型融入束搜索解码以减少偏离目标翻译的通用解码算法。LiBS是一种推理阶段流程，与NMT模型无关，且无需任何额外的平行数据。实验结果表明，我们提出的LiBS算法在WMT和OPUS数据集上平均分别提升1.1 BLEU和0.9 BLEU，并将偏离目标率分别从22.9%降至7.7%、从65.8%降至25.3%。
英文摘要：
Beam search decoding is the de-facto method for decoding auto-regressive Neural Machine Translation (NMT) models, including multilingual NMT where the target language is specified as an input. However, decoding multilingual NMT models commonly produces ``off-target'' translations -- yielding translation outputs not in the intended language. In this paper, we first conduct an error analysis of off-target translations for a strong multilingual NMT model and identify how these decodings are produced during beam search. We then propose Language-informed Beam Search (LiBS), a general decoding algorithm incorporating an off-the-shelf Language Identification (LiD) model into beam search decoding to reduce off-target translations. LiBS is an inference-time procedure that is NMT-model agnostic and does not require any additional parallel data. Results show that our proposed LiBS algorithm on average improves +1.1 BLEU and +0.9 BLEU on WMT and OPUS datasets, and reduces off-target rates from 22.9\% to 7.7\% and 65.8\% to 25.3\% respectively.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
