Click on the titles to view the abstracts.
| Date | Speaker | Title |
| 17 July 2011 | Qiang Li | Smoothing for Statistical Machine Translation Time: Morning Location: The Main Building 410 Abstract: This report discusses different strategies for smoothing the phrasetable in Statistical MT, and gives results over a range of translation settings. In this paper, it shows that any type of smoothing is a better idea than the relative-frequency estimates that are often used. |
| 16 June 2011 | Ji Ma | Introduction to Chain Structure
Conditional Rondom Field Time: Morning Location: The Main Building 410 Abstract: This report introduce the model of Conditional Random Field. The report contains three part: Introduce CRF, Parameter Estimation, Inference. At the end of the report, a c++ implementation of chain structure CRF is presented which aims to help you gain a better understanding of the model. |
| 23 May 2011 | Ji Ma | Using Similarity-Based
Method to Improve SMT System Time: Morning Location: The Main Building 410 Abstract: This is a task-oriented method. The basic idea is that if a sentence occurs both in training set and test set, then we can directly use its reference as the translation. |
| 23 May 2011 | Muhua Zhu | Better Automatic Treebank
Conversion Using A Feature-Based Approach Time: Morning Location: The Main Building 410 Abstract: For the task of automatic treebank conversion, this paper presents a feature-based approach which encodes bracketing structures in a treebank into features to guide the conversion of this treebank to a different standard. Experiments on two Chinese treebanks show that our approach improves conversion accuracy by 1.31% over a strong baseline. |
| 23 May 2011 | Jingbo Zhu | Improving Decoding
Gereralization for Tree-to-String Translation Time: Morning Location: The Main Building 410 Abstract: To address the parse error issue for tree-to-string translation, this paper proposes a similarity-based decoding generation (SDG) solution by reconstructing similar source parse trees for decoding at the decoding time instead of taking multiple source parse trees as input for decoding. Experiments on Chinese-English translation demonstrated that our approach can achieve a significant improvement over the standard method , and has little impact on decoding speed in practice. Our approach is very easy to im-plement, and can be applied to other para-digms such as tree-to-tree models. |