User-Context Aware Explainable Recommendation
Tsinghua University
COM2 Level 2
MR3, COM2-02-26
Abstract:
Personalized recommendation is one of the typical and successful AI applications in past years. Traditional recommender systems generally show recommendations to the user without explanation or with simple collaboration-based reasons, such as "users like/bought this item also like/bought...". Such explanations are not quite convincing. Hence study on better explainable recommendations has been taken as one of the most important and valuable research directions in future recommender systems.
Various user context aware information does help for better explainable recommender system, including users' purchase history, demographic characters, user relationships as well as user generated content information in different platforms or domains. In this talk, besides the overview of our work, I will introduce two recent progresses in detail: jointly modeling social and item visibility for recommendation, and attention-based review-level explainable recommendation, which are published in CIKM2017 and accepted by WWW2018, respectively.
Biodata:
Dr. Min Zhang is an associate professor in the Department of Computer Science & Technology (DCST), Tsinghua University. She specializes in Web search, personalized recommendation, and user modeling. Currently she is also the vice director of State Key Lab. of Intelligent Technology and Systems (central lab), the executive director of MOE-MSRA Key Lab on Media and Search, Tsinghua University. She also serves as Associate Editor for the ACM Transaction of Information Systems (TOIS), Short Paper co-Chair of SIGIR2018, Program co-Chair of WSDM2017 and AIRS2016, Task Organizers of NTCIR Intent and IMine core tasks, and SPC or PC members at SIGIR, WSDM, WWW, IJCAI, KDD, AAAI, ACL, CIKM, etc. She has published more than 100 papers on important international journals and conferences with 2700 citations. She was awarded Beijing Science and Technology Award (First Prize), SIGIR 2017 best student paper awards, etc. She also owns 12 patents.