• ASE Project Proposals
• Abraham Campbell
• Alexey Lastovetsky
• Aonghus Lawlor
• Brett Becker
• Chris Bleakley
• Eleni Mangina
• Fatemeh Golpayegani
• Félix Balado
• Gavin McArdle
• Guenole Silvestre
• Liliana Pasquale
• Mel Ó Cinnéide
• Neil Hurley
  • Parallelisation of Network Alg.
  • Diversifying Rec. Algorithms
  • Personalisation with Privacy
• Performance Engineering
• Rem Collier
• Tony Veale
• Vivek Nallur

Diversifying Recommender Algorithms

Personalised recommendation algorithms use users' profiles of past transactions in order to make recommendations for future transactions. The "People who bought this also bought ..." recommendations that are ubiquitous on platforms such as Amazon and Netflix, are examples of simple recommendation algorithms.  One problem with state-of-the-art recommendation algorithms is that they are focused mainly on accuracy - making sure that the recommendation is relevant to the user - and this sometimes comes at the expense of diversity - the set of recommendations can be very similar to each other. This project will entail the incorporation of diversity into a state-of-the-art recommendation objective, with the goal of increasing the diversity of the recommendation. Recommendation algorithms need to work on large-scale data and part of the focus of this project may be the exploitation of high-performance computing techniques to scale to large data-sets.