Abstract:
”Semantic Web” is the name of the third generation from the evolution of the World Wide Web. The first generation was static in nature and laid the foundation on static pages and content delivery while the second was dynamic with emphasize on the online communities and social interactions. The Semantic Web have as main objective the automation of processes that exist on the Classic Web, allowing machines to execute tasks that currently can be made only by humans, such are those of finding, analyzing, processing, combining and/or sharing of information.
Semantic markups are annotations of Web pages with data about their content that can be understood by the agents on the Web. In order to ensure that different agents have a common understanding of the terms ontologies are used in which are defined the terms and their relations from a domain application , thus establishing a common terminology among agents. In the context of the Semantic Web ontologies play an important role in the sustaining of automated processes (intelligent agents) accessing the information. Particularly, they are used to provide structured vocabularies that describe terms and relations between them, allowing agents to interpret them correcly and unambiguously. Due to the fact that the Semantic Web is in the infancy of its evolution, there is currently a great need for those vocabularies with semantically structured data that describe the significance of web resources, as well as intelligent agents capable to execute the desired operations onto the resources, the most important being those of reasoning.
In the first part of this thesis I set out to begin with an introduction to the domain of discussion, the Semantic Web (chapter 1), followed by a chapter 2 in which I have made a series of discussions, debates, analyses about the foundation stone of the Semantic Web:ontologies, and their representations using logical formalisms, especially Description Logics and First Order Logics. In Chapter 3 I created a reasoning system for the logical knowledge bases that, besides the usual types of knowledges also contained another one called ’negative constraints’, stated in the form of some restrictions on the data from the base. This is the first of its kind system in the literature that realize intensive computations of rules unification on the set of facts with the final goal to find out all the implicit knowledge from the explicitly stated ones, process that in logics is known as ’saturation’.
Part II focuses on the application of Semantic Web technologies into a domain of a great importance for our lives, that is cybersecurity. After a preliminary introduction into the domain that was provided in Chapter 4 , it follows a chapter of discussions and analyses regarding the way are utilized the techniques from Artificial Intelligence and Semantic Web in the construction of intrusion detection systems in order to enhance their performance. Chapters 6 and 7 are the contributions of this second part in the domain in cause. The former proposes an ontology of cybersecurity that is intended as a large model, inside which are defined terms of the domain as well as their relations and their natural language explanations. The ontology is used by a firewall to detect the nature of the occurred situations. Chapter 7 proposes an intrusion detection system in networks of computers that relies on technologies presented throughout this thesis to enhance the detection performance and lower the error rate.
The thesis ends with a chapter in which are drawn the final conclusions upon the work presented, emphasizing the contributions brought to the domain compared to the state of the art from literature, as well as the future directions of my research.