arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1609.07730cs.CL
收藏

用于神经机器翻译的格基循环神经网络编码器

Lattice-Based Recurrent Neural Network Encoders for Neural Machine Translation

Xiamen University(厦门大学);Soochow University(苏州大学);Tsinghua University(清华大学)

Jinsong Su, Zhixing Tan, Deyi Xiong, Rongrong Ji, Xiaodong Shi, Yang Liu

摘要:

神经机器翻译(NMT)严重依赖于词级建模来学习输入句子的语义表示。但是,对于没有自然词分隔符的语言(例如,中文)中输入句子必须首先进行标记化,传统的NMT面临两个问题:1)难以找到用于源句子建模的最佳标记化粒度; 2)1-最佳标记化中的错误可能传播到NMT的编码器。为了解决这些问题,我们提出了基于词格的循环神经网络(RNN)编码器,它将标准RNN推广到词格拓扑结构。所提出的编码器将对多个标记进行压缩编码的词格作为输入,并学习从任意多个输入和先前时间步中的隐藏状态生成新的隐藏状态。因此,基于词格的编码器不仅减轻了标记化错误的负面影响,而且对嵌入输入句子更有表达力和灵活性。中英文翻译实验结果表明,该编码器优于传统的编码器。

英文摘要:

Neural machine translation (NMT) heavily relies on word-level modelling to learn semantic representations of input sentences. However, for languages without natural word delimiters (e.g., Chinese) where input sentences have to be tokenized first, conventional NMT is confronted with two issues: 1) it is difficult to find an optimal tokenization granularity for source sentence modelling, and 2) errors in 1-best tokenizations may propagate to the encoder of NMT. To handle these issues, we propose word-lattice based Recurrent Neural Network (RNN) encoders for NMT, which generalize the standard RNN to word lattice topology. The proposed encoders take as input a word lattice that compactly encodes multiple tokenizations, and learn to generate new hidden states from arbitrarily many inputs and hidden states in preceding time steps. As such, the word-lattice based encoders not only alleviate the negative impact of tokenization errors but also are more expressive and flexible to embed input sentences. Experiment results on Chinese-English translation demonstrate the superiorities of the proposed encoders over the conventional encoder.

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。

↑