IT Ontology and Semantic Technology

IT (information technology) ontology is to be used for analyzing the information technology as well as for enhancing it. Semantic technology is compared with the syntactic one. Ontology plays a backbone for meaning-centered reconfiguration of syntactic structure, which is one aspect of semantic technology. The purpose of use of IT ontology will be categorized into two things: to capture the right information and services for user requests, and on the other hand, to give insights for the future IT with their possible paths by interlinking relations on component classes and instances. Consider question-answering based on ontology to improve the performance of QA. Each question type (e.g., 5W1H) will seek its specific relation from the ontology that has already been acquired from the relevant information resources (e.g., Wikipedia or news articles). The question is whether such relations and related classes are so neutral independent of domain or they are affected by each specific-domain. The first step of ontology learning for question-answering application is to find such neutral relation discovery mechanism and to take care of the special distorted relation-instance mapping when populating on the domain resources. Then, we will consider the domain ontology acquisition by top-down manner from already made similar resources (e.g., domain-specific thesaurus) and also bottom-up manner from the relevant resources. But the already-made resources should be checked against the current available resources for their coverage. Problem is that thesaurus is comprised of classes, not the instances of terms that appear in corpora. They have little coverage over the resources, and even the mapping between classes and instances has not been established yet in this stage. Clustering technology could now filter out the irrelevant mappings. Features of clustering could be improved more accurate by using more semantic ones that have been accumulated during the steps. For example, discovery process based on patterns could be evolved by putting the discovered semantic features into the patterns. Keeping ontology use for question-answering in mind, it is asked for how much the acquired ontology can represent the resources used for acquisition processes. Derived questions are summarized into two about: (1) how such ideal complete ontology could be generated for each specification of use, and (2) how much ontology contributes to the intended problem-solving. The ideal case is to convert all of resources to their corresponding ontology. But if presupposing the gap between the meaning of resources and acquired ontology, a set of raw chunks in resources may be still effective to answer for given questions with some help from acquired ontology or even without resort to them. Definitions of classes and relations in ontology would be manifested through dual structure to supplement the complementary factors between the idealized complete noise-free ontology shape and incomplete error-prone knowledge. In the result, we now confront two problems: how to measure the ontology effectiveness for each situation, and how to compare with the use of ontology for each application and to transform into another shape of ontology depending on application, that could be helped by granularity control and even extended to reconfiguration of knowledge structure. In the result, the intended IT ontology is modularized enough to be compromised later for each purpose of use, and in efficient and effective ways. Still we have to solve definition questions and their translation to ontology forms.