Semantic Technology separates meaning from data. This allows disparate data sources to communicate and interoperate with each other. Semantic Technology makes use of ontologies, which are shared and agreed upon conceptualizations of a domain. Ontologies are both machine-readable and readable by humans. This additional layer of abstraction was developed to increase interoperability between different databases and to make the proliferation of various (historical) data formats manageable. When applied to the Web, we call this concept the Semantic Web (coined by Tim Berners-Lee). The Semantic Web refers to the ‘web of data’, which is considered to be the next generation of the Web. In the Semantic Web machines can understand the meaning of data. This allows for more efficiency as machines rather than humans can perform computations regarding content, relevancy, relations, etc.
Semantic technologies and their principles are widely used in industry and academia today, also in conjunction with other technologies such as data mining, software agents, recommender algorithms, search technologies, social networks etc. At DEBII, we study and develop semantic technology and ontologies for business domains, including but not limited to:
- Ontology engineering, design and implementation
- Onto-system and onto-servers
- Ontology, sub-ontology and commitment
- Ontology merging and alignment
- Ontology learning and evolution
- Ontology presentation tools and applications
- Semantic web and web semantics
- Data semantics and business semantics
Social Network Semantics
Due to the nature of the ontology as a passive structure, as are many ontologies that exist including the software engineering ontology on the web, it is effective in providing support if the end-users know exactly what they are looking for. Frequently, a person addressing the ontology may be trying to resolve an issue and may not be able to translate it into the exact concepts and relationships she/he needs in order to access information in the ontology. What end-users need is an active support. The means of providing such active support to end-users and making it easier for them to find the information they need, and providing them with meaningful output, is the issue to be addressed in this project. We aim to develop a social network-based recommender approach to provide active support and recommendations to remote software engineers. The software engineers will be able to effectively and efficiently access and share evolving software engineering knowledge when carrying out software development. This approach integrates the existing software engineering ontology and recommender approach through the utilization of social network agents. In particular, the existing software engineering ontology enables an active ecology of social network agents to convey, consume and act on project information (semi-) autonomously, according to explicit software engineering domain knowledge. Recommendation techniques are used to convey useful project information and recommend tentative solution(s) for project issues that have been raised by team members.,/p>
The aim is to apply the developed tree mining algorithms to enable ontology building through the matching of existing knowledge representations from the same domain. The main problem to be addressed in this process is to find semantically correct matches among the concept terms in heterogeneous knowledge representations. We will initially avoid considering concept labels as a guide for the formation of candidate mappings but rather use the structural information in which concepts occur in a particular knowledge representation. Taking the structural position of the concept term nodes is, to a certain extent, a promising approach for considering the context in which the concept terms are used. Taking context into consideration is one of the main difficulties in existing approaches. As opposed to matching concepts based upon label comparison, taking the structural aspects into account will indicate possible complex matches (i.e. cases where a concept term in one knowledge representation maps to multiple concept terms in another knowledge representation). The relations considered are limited to the subsumption relations implied by the concept hierarchy or taxonomy. In this respect, the two main problems considered are the matching of knowledge representations at the conceptual and structural levels. Once efficient graph mining approaches have been developed, a similar idea will be applied for obtaining a graph-structured ontology through matching of graph-structured, heterogeneous, knowledge representations of the same domain.
Defeasible Argumentative Reasoning
The aim of this research is to propose and validate a framework for carrying out defeasible argumentative reasoning in semantic web applications. Web is a source of huge amount of data and semantic web efforts are targeted towards making web data contents machines understandable. This will have significant impact on the way information is exchanged and businesses are conducted. As ontology layer of semantic web has got enough maturity (i.e. standards like RDF, RDFs, OWL, OWL 2) the next step is to work on logic layer for development of advance reasoning capabilities for knowledge extraction and efficient decision making. Adding logic to web means use of rules to make inferences. Rules are a way to express business processes, policies, contracts etc but most of the studies have focused on use of monotonic logics in layered development of semantic web which provides no mechanism for representing mechanism for representing incomplete information and handling of contradictory information. These limitations are inherited in Description Logics being subset of predicate logic. Defeasible logic programming is based on nonmonotonic logic and has been used in software agents for carrying out goal driven defeasible reasoning. Defeasible reasoning is a rule based approach to perform reasoning on incomplete, inconsistent and uncertain information and priorities are used to resolve conflicts among rules. Semantic web is source of defeasible knowledge as its open by nature and subject to inconsistencies deriving from multiple sources; therefore, it is not possible to define priorities in advance among conflicting rules. Additionally, quantitative approaches for reasoning on semantic web are also criticized for their in ability to generate easy to understand and logically clear result. We are interested in to exploiting the power of defeasible logic and argumentation for data driven reasoning on semantic web by identifying the issues involved in mapping of RDF/OWL ontologies to defeasible logic programming, how to carry out argumentative reasoning on semantic web, how DeLP rules can be shared on web and how tractable, customizable results can be represented to user.
This research aims to develop a framework to particularly assist Software Engineering Ontology evolution and management by using a semantic wiki based approach. Software Engineering Ontology encompasses common shareable software engineering knowledge and software engineering concepts, how and why they are related. However, most existing ontologies including the Software Engineering Ontology are derived from a single perspective which can be complicated for users, have a lack of maintainability, and become obsolete and impracticable ontologies. Additionally, the knowledge encoded in the ontologies is not static but should evolve. To overcome these impediments to the Ontology Evolution, this research will use a semantic wiki based approach which provides an environment for discussion and formalizing ontology issues, supports ontology evolution, and assist maintaining the versions of ontology.
Ontology Evolution in Oil and Gas
Ontology evolution is one of the main issues that Ontology users face today. The issues include lack of communication between offshore and onshore staff in the development and evolving ontology in oil and gas domain. Additionally, a lot of work for ontology engineers involves updating the ontology manually. The process of evolving the ontology requires work from the domain experts. There has not been a cost effective communication method to discuss and agree upon the changes. In this project, the use of a lightweight community-driven approach is developed in order to enhance the communication on the ontology evolution. This will enable a more efficient and effective method to communicate and take decisions in a better way.
Ontology based Knowledge Sharing Measurement
Ontology provides a shared semantically domain knowledge in a declarative formalism. It specifies consensual knowledge accepted by a community. However meaning and understanding of concepts in ontologies vary in different communities. Determining the semantic similarity or difference of two ontologies is vital to knowledge transferability. How ontology similarity relates to efficiency of knowledge transferability is focused. Complexity of the different knowledge between the two ontologies is fundamental to knowledge sharing. Sharing of knowledge will be efficient and effective if knowledge senders and knowledge receivers having similar understanding of concepts in ontologies and/or the new knowledge is not complicated. In this project two key variables of knowledge sharing i.e. transferability and complexity and how the two variables relate to efficiency of knowledge sharing are focused. A numeric measurement of the transferability between two ontologies is projected. As well as numeric measurement of the complexity of the ontology difference is projected.