Overview#

I am a computer scientist with experience in data science, semantic web technologies, and decision-making AI systems. My work addresses real-world challenges by creating ontologies, building knowledge graphs, and applying automated reasoning to achieve practical solutions.

Research Visit

One of the key projects is the Forest Observatory Ontology (FOO) (ontology.forest-observatory.cardiff.ac.uk), developed with input from domain experts and wildlife data provided by the Danau Girang Field Centre (DGFC) (cardiff.ac.uk/danau-girang-field-centre), a research facility in the Forest of Sabah, Malaysian Borneo. FOO brings together diverse wildlife datasets into an ontology-based knowledge graph as shown in the diagram below.

System Design#

System Design

This knowledge graph was used in training deep learning models and enabling semantic reasoning. Using historical GPS sensor data collected from collars fitted around elephants’ necks, I trained a deep learning model to predict their movements with 99.04% accuracy, outperforming traditional methods such as linear regression (90.95%) and vector autoregression (91.64%). Semantic reasoning rules were also applied to predict potential poaching incidents.

I generalised my semantic data integration approach to a different domain, specifically IoT data marketplaces. This approach allowed data consumers, instead of purchasing sensor datasets in bulk, to buy only the specific data needed for tasks such as training AI models. For this project, I developed an ontology with input from domain experts and populated it with data from six heterogeneous sensors, where each sensor maintained its own knowledge graph. Semantic reasoning rules were applied to these knowledge graphs to address practical use cases.

Connect with me