Data Semantics (ECS7028U/ECS7028P)

While data has become a valuable asset across industries in recent years, organisations increasingly realise that having large amounts of data is not sufficient to derive value from it. Data needs to be clean, consistent, preferably interconnected and associated with clear semantics. This enables data scientists and business analysts to focus on extracting useful insights from vast amounts of data, especially in the world of social media. Examples of semantic data models include knowledge graphs, ontologies and taxonomies that have been developed in the data and artificial intelligence world for the past decade. The goal of these models is to capture the meaning of data in an explicit and shareable way, and to facilitate data-driven applications. The popularity of these models has increased substantially through the development of knowledge driven search at internet companies, the development of the Semantic Web, social networks, as well as media sharing and streaming platforms. This module will teach students fundamental principles of semantic data modelling though discussing applications related to the Semantic Web and Semantic Media. This includes logic based data modelling principles which strike a good balance between tractability and usability, and data modelling languages such as the Ontology Web Language (OWL) and the graph-based SPARQL query language. These allow automated processing and reasoning over data and facilitate the use of AI techniques in tasks such as search and recommendation. The module introduces implementations and applications of data semantics to a broad range of content types including music and media. Topics include XML, JSON, semantic modelling, Predicate and Description Logics, the RDF model and databases, OWL2, SPARQL and ontology design, as well as applications and ontologies for specific domains including text, image, audio and multimedia.