Improvement on belief network framework for natural language understanding.

Mok, Oi Yan.Thesis (M.Phil.)--Chinese University of Hong Kong, 2003.Includes bibliographical references (leaves 94-99).Abstracts in English and Chinese.Chapter 1 --- Introduction --- p.1Chapter 1.1 --- Overview --- p.1Chapter 1.2 --- Thesis Goals --- p.3Chapter 1.3 --- Thesis Outline --- p.4Chapter 2 --- Background --- p.5Chapter 2.1 --- Natural Language Understanding --- p.5Chapter 2.1.1 --- Rule-based Approaches --- p.7Chapter 2.1.2 --- Phrase-spotting Approaches --- p.8Chapter 2.1.3 --- Stochastic Approaches --- p.9Chapter 2.2 --- Belief Network Framework - the N Binary Formulation --- p.11Chapter 2.2.1 --- Introduction of Belief Network --- p.11Chapter 2.2.2 --- The N Binary Formulation --- p.13Chapter 2.2.3 --- Semantic Tagging --- p.13Chapter 2.2.4 --- Belief Networks Development --- p.14Chapter 2.2.5 --- Goal Inference --- p.15Chapter 2.2.6 --- Potential Problems --- p.16Chapter 2.3 --- The ATIS Domain --- p.17Chapter 2.4 --- Chapter Summary --- p.19Chapter 3 --- Belief Network Framework - the One N-ary Formulation --- p.21Chapter 3.1 --- The One N-ary Formulation --- p.22Chapter 3.2 --- Belief Network Development --- p.23Chapter 3.3 --- Goal Inference --- p.24Chapter 3.3.1 --- Multiple Selection Strategy --- p.25Chapter 3.3.2 --- Maximum Selection Strategy --- p.26Chapter 3.4 --- Advantages of the One N-ary Formulation --- p.27Chapter 3.5 --- Chapter Summary --- p.29Chapter 4 --- Evaluation on the N Binary and the One N-ary Formula- tions --- p.30Chapter 4.1 --- Evaluation Metrics --- p.31Chapter 4.1.1 --- Accuracy Measure --- p.32Chapter 4.1.2 --- Macro-Averaging --- p.32Chapter 4.1.3 --- Micro-Averaging --- p.35Chapter 4.2 --- Experiments --- p.35Chapter 4.2.1 --- Network Dimensions --- p.38Chapter 4.2.2 --- Thresholds --- p.39Chapter 4.2.3 --- Overall Goal Identification --- p.43Chapter 4.2.4 --- Out-Of-Domain Rejection --- p.65Chapter 4.2.5 --- Multiple Goal Identification --- p.67Chapter 4.2.6 --- Computation --- p.68Chapter 4.3 --- Chapter Summary --- p.70Chapter 5 --- Portability to Chinese --- p.72Chapter 5.1 --- The Chinese ATIS Domain --- p.72Chapter 5.1.1 --- Word Tokenization and Parsing --- p.73Chapter 5.2 --- Experiments --- p.74Chapter 5.2.1 --- Network Dimension --- p.76Chapter 5.2.2 --- Overall Goal Identification --- p.77Chapter 5.2.3 --- Out-Of-Domain Rejection --- p.83Chapter 5.2.4 --- Multiple Goal Identification --- p.86Chapter 5.3 --- Chapter Summary --- p.88Chapter 6 --- Conclusions --- p.39Chapter 6.1 --- Summary --- p.89Chapter 6.2 --- Contributions --- p.91Chapter 6.3 --- Future Work --- p.92Bibliography --- p.94Chapter A --- The Communicative Goals --- p.100Chapter B --- Distribution of the Communicative Goals --- p.101Chapter C --- The Hand-Designed Grammar Rules --- p.103Chapter D --- The Selected Concepts for each Belief Network --- p.115Chapter E --- The Recalls and Precisions of the Goal Identifiers in Macro- Averaging --- p.12