Research On The Intelligent Information Processing Based On The Network-Computing Mode

The intelligent information processing or knowledge processing in the network-computing mode is a new and hot researching field of intelligent information processing or Knowledge Engineering (KE). Distributed intelligent applications (DIA) is an important part of this field which generating by integrating network-computing technique and Artificial Intelligence (AI) technique. Around the hot topic, this thesis presents some research results from two respects of theory and applications. Main research lies in methods and techniques of knowledge acquisition and processing.1. Language Knowledge Acquisition Mode & AlgorithmsThis thesis makes a systematic analysis of the structure, implementation, and applications for computing language modeling. Some improved models and the corresponding implementation methods and algorithms are proposed. In order to overcome the disadvantages of traditional rule-based word-segmenting methods, a new method for segmenting the input Chinese language text sentence into words is presented, which consists of a character-based N-gram model and a self-organizing words algorithm based on machine-learning. The effectiveness has been confirmed by evaluation experiments using the closed texts and open texts corpus. Traditional speech recognition system has an intrinsic defect that, commonly only use the acoustic model of speech and unable to use non-acoustic knowledge of language to recognize speech. A Chinese speech recognition method based on character-based N-gram model is proposed in this thesis. we give the estimation formulas for the parameters of the model and the algorithms for training and recognition. The analysis and experiment shows that introducing SLM reduces the difficulty and improves the performance for Chinese speech recognition.2. A Fuzzy Vector-Correlative Information Search ModelBased on the vector-correlative model (VCM), that is one of the popular information search models. This thesis inducted fuzzy sets theory and approach for to construct search models, and presented a new model, that is Fuzzy Vector-correlative Model (FVCM). So, its theory structure and user's interest has been dimming improved. We have show with some fringe experiments that the new model has overcome some intrinsic defects of the exciting ones, and has raised the precision and recall of information search system.