Granular computing for the design and implementation of intelligent information systems

信息和知识粒化在以人为中心的信息处理和问题解决中起着核心作用。任何真正智能的信息系统都必须能够在多个粒度层次上处理信息和推理。颗粒计算是致力于该主题研究的新兴领域。研究人员面临的一个巨大挑战是缺乏一个被广泛接受的共同理解和统一的框架。在提出的方案中，我不仅寻求一个统一的颗粒计算视图，而且将其原理应用于实现智能信息系统。研究方案分为三个部分：构建概念框架、构建统一模型、基于框架和模型实现原型信息系统。通过借鉴许多科学分支的成果，将在三个层次上充分发展颗粒计算的概念框架。颗粒计算可以被看作是哲学层面的结构化思维，实践层面的结构化问题解决，以及计算层面的结构化信息处理。这些观点自然地对应于信息系统的概念理解、逻辑设计和软件实现。综合现有研究和相关理论的零散结果，建立统一的颗粒计算模型。框架和模型的价值将通过实现原型系统来评估，例如信息检索支持系统、研究支持系统和智能数据分析系统。

Information and knowledge granulation plays a central role in human-centered information processing and problem solving. Any information system of true intelligence must be able to process information and reason at multiple levels of granularity. Granular computing is an emerging field devoted to the study of this topic. A great challenge faced by researchers is the lack of a well-accepted common understanding and a unified framework. In the proposed program, I not only search for a unified view of granular computing, but also apply its principles to implement intelligent information systems. The research program has three parts: building a conceptual framework, constructing a unified model, and implementing prototype information systems based on the framework and the model. A conceptual framework of granular computing will be fully developed at three levels by drawing results from many branches of science. Granular computing can be viewed as structured thinking at the philosophical level, structured problem solving at the practical level, and structured information processing at the computational level. These views correspond naturally to the conceptual understanding, logical design and software implementation of information systems. A unified model of granular computing will be built by synthesizing scattered results from existing studies and related theories. The value of the framework and the model will be evaluated by implementing prototype systems, such as information retrieval support systems, research support systems, and intelligent data analysis systems.