Does String-Based Neural MT Learn Source Syntax?

@inproceedings{Shi2016DoesSN,
  title={Does String-Based Neural MT Learn Source Syntax?},
  author={Xing Shi and I. Padhi and Kevin Knight},
  booktitle={EMNLP},
  year={2016}
}
  • Xing Shi, I. Padhi, Kevin Knight
  • Published in EMNLP 2016
  • Computer Science
  • We investigate whether a neural, encoderdecoder translation system learns syntactic information on the source side as a by-product of training. We propose two methods to detect whether the encoder has learned local and global source syntax. A fine-grained analysis of the syntactic structure learned by the encoder reveals which kinds of syntax are learned and which are missing. 
    222 Citations

    Figures, Tables, and Topics from this paper

    Explore Further: Topics Discussed in This Paper

    Multi-Source Syntactic Neural Machine Translation
    • 15
    • PDF
    Towards String-To-Tree Neural Machine Translation
    • 118
    • PDF
    Syntactic Structure from Deep Learning
    • 9
    • PDF
    Extracting Syntactic Trees from Transformer Encoder Self-Attentions
    • 20
    • PDF
    Encodings of Source Syntax: Similarities in NMT Representations Across Target Languages
    • Highly Influenced
    • PDF
    What do Neural Machine Translation Models Learn about Morphology?
    • 201
    • Highly Influenced
    • PDF
    Modeling Source Syntax for Neural Machine Translation
    • 100
    • PDF

    References

    SHOWING 1-10 OF 32 REFERENCES
    What Can Syntax-Based MT Learn from Phrase-Based MT?
    • 111
    • PDF
    Skip-Thought Vectors
    • 1,602
    • PDF
    Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
    • 10,532
    • PDF
    Grammar as a Foreign Language
    • 761
    • PDF
    What's in a translation rule?
    • 543
    • PDF
    Scalable Inference and Training of Context-Rich Syntactic Translation Models
    • 477
    • PDF
    A Discriminative Model for Tree-to-Tree Translation
    • 98
    • PDF
    A Convolutional Neural Network for Modelling Sentences
    • 2,519
    • PDF
    An Empirical Examination of Challenges in Chinese Parsing
    • 19
    • PDF
    Statistical Phrase-Based Translation
    • 3,579
    • PDF