Liangyou Li, ESR8

Location: Dublin City University, Ireland

Project Title: SMT/TM Combination

Project Description: 

I started work on the Expert Project in June 2013 as an early stage researcher. My research focuses on the combination of statistical machine translation (SMT) and translation memory (TM) and syntax-based machine translation.

In the past two years, I proposed a discriminative framework which can integrate TM into SMT by incorporating TM-related feature functions. And with the capacity of handling a large amount of features in the discriminative framework, we further proposed a method to use multiple fuzzy matches efficiently which brings more feature functions.

In addition, on a single dependency-based SMT, I proposed to easily implement this model in Moses by transformation and decompose dependency structures to improve this model. More recently, based on an edge replacement grammar, I have proposed to use a synchronous graph-to-string grammar for dependency-to-string translation. This new model has shown further improvements.

I am currently working on an in-depth study into the impact that SMT/TM combination, the improved dependency-based model and their further combination can have on post-editing efficiency in a real world setting, using data provided by Hermes, one of the commercial partners in the EXPERT project. 

Research Interests: Machine Learning, Corpus-Based Machine Translation, Natural Language Processing

Publication list

  1. Liangyou Li, Jian Zhang, Chris Hokamp, Xiaofeng Wu, Xiaojun Zhang and Qun Liu (2013) The CNGL MT System for CWMT’2013, in Proceedings of China Workshop of Machine Translation (CWMT 2013), 31 October, Kunming, China

  2. Liangyou Li, Xiaofeng Wu, Santiago Cortés Vaíllo, Jun Xie, Andy Way, Qun Liu. (2014). The DCU-ICTCAS MT system at WMT 2014 on German-English Translation Task. In Proceedings of the Ninth Workshop on Statistical Machine Translation, pages 136–141, Baltimore, Maryland USA, June 26–27, 2014.

  3. Liangyou Li, Jun Xie, Qun Liu. (2014). Transformation and Decomposition for Efficiently Implementing and Improving Dependency-to-String Model in Moses. In Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation, pages 122–131, October 25, 2014, Doha, Qatar.

    http://aclweb.org/anthology/W14-4014

  4. Liangyou Li, Andy Way, Qun Liu. (2014). A Discriminative Framework of Integrating Translation Memory Features into SMT. In Proceedings of the 11th Conference of the Association for Machine Translation in the Americas, Vol. 1: MT Researchers Track. Pages 249-260. Vancouver, BC, Canada. 2014.

  5. Liangyou Li, Hui Yu, Qun Liu. (2015). MT Tuning on RED: A Dependency-Based Evaluation Metric. In Proceedings of the Tenth Workshop on Statistical Machine Translation, pages 428-433, Lisboa, Portugal, 17-18 September.

    http://www.statmt.org/wmt15/pdf/WMT55.pdf

  6. Liangyou Li, Andy Way, Qun Liu. (2015). Dependency Graph-to-String Translation. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 33-43, Lisbon, Portugal, 17-21 September

  7. Liangyou Li, Andy Way and Qun Liu (2016) Phrase-Level Combination of SMT and TM Using Constrained Word Lattice. In Proceedings of ACL 2016, 7-12 August, Berlin, Germany

  8. Liangyou Li, Carla Parra Escartín and Qun Liu (2016) Combining Translation Memories and Syntax-Based SMT: Experiments with real industrial data, In Proceedings of EAMT 2016, 30 May – 1 June, 2016, Riga, Latvia

  9. Liangyou Li , Andy Way, Qun Liu. (2016). Graph-Based Translation Via Graph Segmentation. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Berlin, Germany.

  10. Jian Zhang, Liangyou Li , Andy Way, Qun Liu. (2016). Topic-Informed Neural Machine Translation. In Proceedings of the 26th International Conference on Computational Linguistics. Osaka, Japan. (To appear)

  11. Li, Liangyou; Parra Escartín, Carla; Way, Andy; and Liu, Qun (2016) "Combining Translation Memories and Statistical Machine Translation Using Sparse Features", to appear in Machine Translation Journal. Special Issue: NLP for Translation Memories.

  12. Jian Zhang, Liangyou Li, Andy Way, Qun Liu. (2014). A Probabilistic Feature-Based Fill-up for SMT. In Proceedings of the 11th Conference of the Association for Machine Translation in the Americas, Vol. 1: MT Researchers Track. Pages 96-109. Vancouver, BC, Canada.

Resources