UAM Subproject

The UAM subproject is conducted by the Information Retrieval Group at UAM and focuses on the incorporation of novelty, diversity, subjectivity, context, and time in Information Retrieval (IR) paradigms and models, in order to enhance the quality, effectiveness, and user satisfaction with retrieval systems. The subproject addresses practical, methodological and theoretical research goals, aiming at the definition of enhanced IR and recommendation methods, models and evaluation methodologies that help overcome specific limitations in the field.

The research goals of the subproject include:

  • Definition of new retrieval models and theories that incorporate novelty, diversity, subjectivity, context and time dimensions, along with relevance, as integral elements of the effective information value and user satisfaction. This involves overcoming classic IR paradigms and a revision of traditional assumptions on relevance, such as being topical, independent, consistent, and stable.
  • Definition of new IR and recommendation methods and techniques that enhance the novelty and diversity of retrieved information, while achieving a satisfactory trade-off with accuracy and relevance for user interests.
  • Definition of new methods and paradigms that represent and leverage contextual information to enhance the effectiveness of retrieval and recommendation, focusing on the social user context and the temporal dimension.
  • Investigation of the relations of novelty and diversity to context, personalization and time in IR and recommender systems.
  • Research of enhanced recommender system hybridization methods, and the integration of heterogeneous sources of evidence, towards the enhancement of recommendation accuracy and diversity.
  • Definition of new, formally grounded effectiveness metrics that incorporate novelty and diversity in the evaluation of IR and recommender systems.
  • Revision of recommender system evaluation techniques and design of new methodologies that better capture the effectiveness of recommender systems in real applications. In particular, the sound adaptation of IR evaluation methodologies to recommendation tasks will be researched.
  • Connection and unification of IR and recommender system theories, models, and evaluation methodologies, towards a better understanding of the equivalences and differences between theories and techniques in both fields, identifying and leveraging opportunities for mutual benefit.

Comments are closed.