Home --> Research

Digital data and connected worlds of physical objects, people and devices are rapidly changing the way we interact with our surrounding environments. The cyber-physical and social data and networking technologies have a profound impact on different domains such as healthcare, environmental monitoring, urban systems, and control and management applications, among others. However, these large volumes of heterogeneous data require new methods of communication, aggregation, processing and mining in order to extract efficient and actionable information from the raw data.

My main research goal is to develop intelligent information communication, discovery and retrieval methods for cyber-physical and social systems. I work on machine learning, Internet of Things (IoT), semantic web, services computing, adaptive algorithms, data-centric networking, big data, sream processing and information search and retrieval to solve problems and develop new technologies for the future Internet and Web systems.

The results shown below are mainly from the works of the research team at the University of Surrey and also from the activities in collaborative projects in which some of the works have been developed in collaboration with other partners in the projects.

  • Large-scale Service and Data Discovery

    This research is focused on developed novel methods for large-scale data and service discovery. We have developed a simulation environment for large-scale Internet of Things networks and distributed service platforms. The service discovery method was developed using a probabilistic machine learning approach that is able to identify implicit relations between the service descriptions. This method relies on an unsupervised learning method and can improve the performance of discovery of large-scale semantically annotated services with a significant rate. This probabilistic method outperformed the existing graph based and logical query methods that are used for search and discovery of semantic services in the existing solutions. For the data discovery, we used a probabilistic machine learning approach to index very large-scale semantically described IoT data. We have also developed a compensation mechanism for updating the indexing model which made the solution significantly scalable and computationally effective for dynamic environments such as the IoT.

    • Selected publications
    • Software
      • IoT Configuration Manager and Resource Search and Discovery (open source), this component is a part of the EU Future Internet Test-bed (FI-Lab) (http://www.fi-ware.org/lab/), https://github.com/UniSurreyIoT

      • Simulator for large-scale distributed sensor networks and distributed indexing and discovery (the software is designed based on an industry collaboration and the core algorithm is currently filed for a patent).

  •  Ontology Learning and Semantic Search

    This work focused on automated creation of ontologies from unstructured text. The terminological ontologies created by our proposed method are significantly accurate and can be directly used for semantic search and reasoning purposes. This work also led to further development of a semantic knowledge-base creation and search engine for discovering and ranking scientific papers. Another interesting aspect of the work is being domain independent that makes it applicable to different text libraries and online resources. We also developed a ranking algorithm, called RareRank, that uses the provided knowledge to determine inter-relations between different entities (such as authors and topics and organisations, etc.) to provide more meaningful search results. The evaluation results show that RareRank outperforms Google’s PageRank in searching scientific articles. The proposed method can also rank journals, conferences and researchers according to research topics. Some of the key results and publications from this research are presented in:
      

 
Go back to home page