Deep Transfer Learning for Recommender System

• Goal: Recommender System is playing a key role in providing personalized experience to users of many systems such as E-commerce websites, online news/videos. Despite of significant progress over the past few years, recommender system is still facing the long-standing data sparsity problem. This problem arises due to the fact that each item (i.e., books, movies) are only purchased or rated by a few users, meanwhile the number of items are very large (hundred of thousands). This project aims to address the data sparsity problem in recommender system by means of transfer learning, which is able to transfer knowledge learned from one domain to another related domain. 

• Related publications:

  • Duc Nguyen et al., "On the Transferability of Deep Networks for Recommendation Systems" (ECMLPKDD2020 IAL2020 workshop, Sep. 2020)

  • Hao Niu, Duc Nguyen, "Multi-source Transfer Learning for Human Activity Recognition in Smart Homes",  (IEEE SMARTCOMP, Sep. 2020)