Semantic Conflict Detection And Its Implementation Based On Implication Reasoning

In the field of complex product design, knowledge entry and editing process is usually completed by multiple online collaborators. The Collaboration can improve the efficiency of knowledge sharing, at the same time; it also brings the problem of redundancy and conflicts between the knowledge. This paper presents a framework and its implementation of the semantic conflict detection. This method can identify the contradictions and conflicts that may be inherent between the two natural language texts. And we use this method in the knowledge management system to ensure consistency and correctness of the content in the knowledge Base.Inspired by the human characteristics in the process of natural language understanding, we extract semantic information units expressed in the sentence by the method of lexical, syntax and semantic analysis from the sentence. An algorithm that calculates the semantic similarity between two sentences by greedy matching the semantic units of these two sentences. In this manner, the sentences with low similarity will be excluded. On this basis, we classify and modeling the common conflict. According to the characteristics of a variety of conflict classification, combined with a large knowledge base and rules, we analyze and identify the semantic conflict between the two sentences. Finally, we analysis and explore the techniques that may be used in the recognition of conflict.We apply this semantic conflict identification method in a complex-product-design knowledge management system as a service, which is used in the knowledge entry, editing and evaluation process.