Context-Awareness in Information Retrieval and Recommender Systems
Information retrieval (IR) and recommender systems (RSs) are two important applications and research topics in Web intelligence. They have been developed in several decades in order to alleviate the information overload problem. More specifically, IR delivers a list of documents related to user’s explicit queries, while RS is able to assist decision making by providing a list of suggested items that are tailored to users’ tastes. In recent years, researchers started to realize the importance of context – it is crucial and necessary to take context information (e.g., time, location, companion, emotions, occasion, topics, etc) into consideration. For example, in IR, the judgement of relevance may vary from contexts to contexts, while users’ tastes may be different in RS too. A popular example is that, user’s choice on movies may be different if the user is going to watch the movie with partner rather than with children. Similarly, a user may choose a different restaurant if he or she is going to have a formal business dinner rather than a quick lunch. This short tutorial is going to provide a comprehensive survey about the context-awareness in IR and RS, including the background of IR and RS, definition of contexts, context acquisition, motivations behind in context-aware IR and RS, the approaches of taking contexts into consideration, as well as challenges and promising research in future work.
Ontology Learning and Population from Text Background
Ontologies are used by modern knowledge systems for representing and sharing knowledge about an application domain. Supporting semantic processing, knowledge systems allow for more precise information interpretation, thus providing greater usability and effectiveness than traditional information systems. Manual construction of ontologies by domain experts and knowledge engineers is a costly task, therefore, automatic and/or semi-automatic approaches to their development are needed, a field of research that is usually referred to as ontology learning and population. This short course will focus on discussing the main problems and corresponding solutions for the automatic and/or semi-automatic acquisition of each one of the components of an ontology (classes, properties, taxonomic and non-taxonomic relationships, axioms and instances) from textual resources, through a common process for the incremental building of ontologies driven by the goals of a particular knowledge system. Applied techniques from the areas of Machine Learning, Natural Language Processing, Information Retrieval, Information Extraction and Ontology Reuse.