Key Duties and Responsibilities
Full-time research fellow on the SICSA Smart Tourism project’s “Living History” and “Smart Beacons”.
- 2013: PhD in Computer Science. Thesis title: Integrating Content and Semantic Representations for Music Recommendation
- 2009: BSc (Hons) in Computer Science. Awarded class prize. Dissertation title: Notational Output from Input Sound Engine
Research Interests / Professional Background
Dr Horsburgh is a Research Fellow within the IDEAS research institute at Robert Gordon University. His research focuses on knowledge representation, which feeds into recommender systems.
Recent projects have a focus on Smart Tourism, and how personalised location-aware delivery of information can be provided for tourists. Previous work has investigated knowledge refinement in a textiles exploration system, and recommending lessons learned for the oil industry using a case-based approach.
Dr Horsburgh's PhD thesis looked at knowledge representation in music recommender systems, and developed novel representations to overcome current challenges within the field. The three main contributions of this work were:
- An improved description of musical audio texture, MFS
- The development of hybrid texture-tag representations that reduce cold-start, and increase discovery of novel tracks in a music recommender system
- A novel approach to combining implicit and explicit feedback to evaluate music recommender systems
Additional Information / Media Work / Funding
- Scottish Funding Council/Smart Tourism funded innovation project “Smart Beacons” 2012. £72k
- Horsburgh, B. (2013) Integrating content and semantic representations for music recommendation. Ph.D Thesis. Robert Gordon University: U.K.
- Horsburgh, B., Craw, S., Williams, D., Burnett, S., Morrison, K., & Martin, S. (2013). User Perceptions of Relevance and Its Effect on Retrieval in a Smart Textile Archive. In Case-Based Reasoning Research and Development, Lecture Notes in Computer Science Volume 7969 (pp. 149-163). Springer Berlin Heidelberg.
- Horsburgh, B., Craw, S. & Massie, S. (2012). Music-inspired texture representation, Proceedings of the National Conference on Artificial Intelligence (AAAI), pp. 52-58.
- Horsburgh, B., Craw, S. & Massie, S. (2012a). Cold-start music recommendation using a hybrid representation, Digital Economy All Hands Conference.
- Horsburgh, B., Craw, S., Massie, S. & Boswell, R. (2011). Finding the hidden gems: Recommending untagged music, Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence (IJCAI), pp. 2256-2261.