• Utilizing side information about items for user modeling and recommending including structured sources, e.g., DBpedia, Linked Open Data, BabelNet, Wikidata, textual sources, e.g., Wikipedia or User-Generated Content like tags, reviews, and comments multimedia (“low-level”) features, e.g., videos or musical signals
  • Approaches that rely on a semantic (deep) understanding of items and their features based, e.g., on formal ontologies
  • Applying deep learning methods to model item features
  • Leveraging rich item representations for more effective user modeling and recommendation
  • Using side information about items to increase recommendation quality in terms of novelty, diversity, or serendipity
  • Using side information about items to explain recommendations to users
  • Leveraging side information and external sources for cross-lingual recommendations
  • Using side information about items for transparent user modeling compliant with the General Data Protection Regulation
  • Novel applications areas for recommender systems (e.g., music or news recommendation, off-mainstream application areas) based on item side information
  • User studies (e.g., on the user perception of recommendations), field studies, in-depth experimental offline evaluations
  • Methodological aspects (evaluation protocols, metrics, and data sets)