论文标题
神经机器翻译的质量意识解码
Quality-Aware Decoding for Neural Machine Translation
论文作者
论文摘要
尽管过去几年的机器翻译质量估计和评估取得了进展，但在神经机器翻译（NMT）中进行解码大多忽略了这一点，并围绕根据模型（地图解码）找到最可能的翻译，并与光束搜索近似。在本文中，我们通过各种推理方法（例如$ n $ best Reranking and Timumimum best Rearanking and Miniumimim bases风险危险解码）利用了最近的突破，并提出了NMT的质量意识解码。我们对四个数据集和两个模型类别的各种可能的候选生成和排名方法进行了广泛的比较，发现质量意识的解码始终优于基于图的基于图的解码，这是根据最先进的自动指标（Comet和Bleurt）和人类评估。我们的代码可在https://github.com/deep-pin/qaware-decode上找到。
Despite the progress in machine translation quality estimation and evaluation in the last years, decoding in neural machine translation (NMT) is mostly oblivious to this and centers around finding the most probable translation according to the model (MAP decoding), approximated with beam search. In this paper, we bring together these two lines of research and propose quality-aware decoding for NMT, by leveraging recent breakthroughs in reference-free and reference-based MT evaluation through various inference methods like $N$-best reranking and minimum Bayes risk decoding. We perform an extensive comparison of various possible candidate generation and ranking methods across four datasets and two model classes and find that quality-aware decoding consistently outperforms MAP-based decoding according both to state-of-the-art automatic metrics (COMET and BLEURT) and to human assessments. Our code is available at https://github.com/deep-spin/qaware-decode.
