The research described in this thesis aims at the design, implementation and evalu- ation of a natural-language processing system that is used as part of an information retrieval system. Specically, I will focus on the development of a system that per- forms robust syntactic analysis of scientic texts written in English. Robustness is regarded as the property of a computer system that, no matter what input, the sys- tem will not fail in producing an appropriate output, preferably an output that can be guaranteed to be correct.

Document Analysis and Understanding (DAU) is a complex AI application with high industrial impact. For the increasing demands upon the bandwidth and quality of the analysis it is crucial to enable different analysis modules to collaborate. For making collaboration possible, we first examine the question whether there exists a common ontological basis which can serve as a platform for communication of different DAU modules. Once communication is enabled, we investigate the second question, how DAU modules originally designed as stand-alone systems must be modified in order to benefit from collaboration with others.

This paper proposes a novel method to refine the grammars in parsing by utiliz- ing semantic knowledge from HowNet. Based on the hierarchical state-split ap- proach, which can refine grammars au- tomatically in a data-driven manner, this study introduces semantic knowledge into the splitting process at two steps. Firstly, each part-of-speech node will be anno- tated with a semantic tag of its termi- nal word. These new tags generated in this step are semantic-related, which can provide a good start for splitting. Sec- ondly, a knowledge-based criterion is used to supervise the hierarchical splitting of these semantic-related tags, which can al- leviate overfitting. The experiments are carried out on both Chinese and English Penn Treebank show that the refined gram- mars with semantic knowledge can im- prove parsing performance significantly. Especially with respect to Chinese, our parser achieves an F 1 score of 87.5%, which is the best published result we are aware of