Abstract:This paper proposes an adaptive hierarchical vector quantization method for image coding. First, a basic codebook and the codes are generated adaptively. Then the codes are used to form the second layer codebook. Finally, the codes are coded by the second layer codebook and form a multilayer code sequence. This hierarchical method makes use of the structural features and correlations in larger regions and thus remove much redundance in the codes obtained by a single layer VQ method. A much lower bit rate can beachieved without causing any more distortion and more computation.

Abstract:A lot of useful and valuable information may be included in expert analytical cases, which is absent in ordinary cases. Based upon detailed observation and analysis analytical cases, the paper presents an algorithm which can learn five kinds of rules from them.

Abstract:Generally, one way to represent temporal information in a logical formalism is by associating "proposition types" with time points or time intervals. It was pointed out that time intervals are difficult to represent the common sense notion of a single event with "gaps" in [1]. However, the temporal relations of two gap -intervals defined in [1] are expressed in a matrix. The computing amount of temporal relations is so large that the method in [1] is not useful. This paper presents a interval-gap-based representation of temporal knowledge. Its temporal relations are stipulated to 20, and its computing amount is in same level to Allen's.

Abstract:A concept of semantic relationship between fuzzy sets is presented in the paper. The concept is not only related to the membership of the fuzzy sets, but also considered releant to the similarity between elements in the domains of the fuzzy sets. So,it can be used to characterize the similarity between fuzzy sets more meticulously than it was done by the various "distance" or "nearness" previously defined in fuzzy mathematics.There fore, it would have a wide range of applications in AI and other areas, such as information retrieval and text classification. Besides, two algorithms to approximately calculate the semantic relationship are given in the paper.

Abstract:Word classification is still an open issue for the linguistic research of Chinese.A fixed-point theory of word classification is proposed in this paper. According to our result,the Distributed Analysis (DA) method for word classification can be used in a process of iteration, so that the "logical circle" critique on DA method does not hold. The ideal classification, i. e., the maximal fixed-point, is computable in the limit. This paper therefore cleared out the misunderstandings on DA method among linguists, and proved the distribution criterion to be scientific.

Abstract:ntelligent Decision Support Systems are developed at an increasing rate. But there exist many problems of integration of AI and DSS, on which we have done some research while developing the intelligent DSS system, NUIDSS. This paper describes its design and implementation, focusing on model description language, knowledge base and inference. A practical decision instance of emerge is also given in the end.

Abstract:n this paper, a kind of Trie structure is described. The insertion and search algorithms on this kind of Trie structure are presented. The time complexity the search algorithm is O(lognK). Comparing to prior work, it is a improvement. This paper also presents the search algorithm on this kind of Trie structure stored into an one dimension array.

Abstract:This paper presents the design considerations and implementation technics of an open system TX, a kernel for distributed database systems developed by the database group of the computer science department at Tsinghua University from 1987 to 1990. TX has played important roles in their further research on distributed database system,knowledge based system, and hypertext database since it was partially implemented. The focuses of this paper are the architecture, data organization, query, processing, data dictionary, and concurrency control of TX.

Abstract:his paper points out that some problems in designing and maintaining are caused with some inherent difficulties in expert system. The paper makes the suggestion of software engineering method about ES, the knowledge specification (KS) about cognitive process, and PLWF model of the process for building ES. The problems could be solved with PLWF model and KS, in the stage of building ES and in the stage of maintaining ES.

Abstract:This paper presents a ADT-based functional language which is the functional component of a hybrid language that the authors designed supporting both the functional programming and object -oriented programming, and describes its operational semantics model: polymorphic λ- calculus + algebraic rewriting system = polymorphic λ- rewriting system. Moreover, the Church - Rosser property and strong normalization property of this model are discussed.