Computational Linguistics Colloquium

Tuesday, 17 April, 16:15
Conference Room, Building C7 4

Hybrid Approach towards Chinese base-NP Chunking at CIP-CASIA

In this talk, I will introduce our work on text chunking at the Chinese Information Processing Group, Institute of Automation, Chinese Academy of Sciences. First of all, I will give a survey and overview of some state-of-art method for noun phrase chunking. Then, particularly I will introduce our hybrid approach for chunking Chinese base noun phrase, which combines SVM (Support Vector Machine) and CRF (Conditional Random Field) model with handcrafted grammar rules and probabilistic criterion from CRF. Successively, our error-driven SVM classifier which was designed to learn the errors found by comparison between TBL (Transformation-based Learning) and CRF chunking classifiers. According to the experiments, our method improved the overall performance of noun phrase recognition. Finally, I propose some future work for shallow parsing.

If you would like to meet with the speaker, please contact Yi Zhang.