Yue Shi

Research Scientist
Machine Learning in Business Integrity, Ads, Marketplace, Pages

Previously I worked at Yahoo! Labs / Yahoo! Research on recommender systems and information retrieval, including (past) projects:
  • Large Scale User Profiling
    • Contribution: Implement end-to-end user profile pipeline for Yahoo! Homepage
    • Impact: +1% lift in the major online metric of Yahoo! Homepage
    • Related Publication: KDD '15
  • Large Scale Factorization Models
    • Contribution: Co-work/Implement on distributed latent factor models built on Map-Reduce
    • Impact: +2% lift in the major online metric of Yahoo! Homepage
    • Related Publication: CIKM '16 (Slides)
  • Large Scale Context-aware Recommendation
    • Contribution: Implement contextual mobile app recommendation for Yahoo! Aviate
    • Impact: +100% lift in the major online metric of the product
  • Social Recommendation
  • Vertical Search and Ranking
  • Video Recommendation
    • Contribution: Implement personalized model of Playlist recommendation in Yahoo! Video Player
    • Impact: +20% lift in key online metrics for Player

I received my PhD degree from Delft University of Technology in 2013, under the supervision of Prof. Alan Hanjalic, Dr. Martha Larson, and Prof. Inald Lagendijk. My PhD research is on Recommender Systems, and my thesis "Ranking and Context-awareness in Recommender Systems" is here.

I obtained my BSc degree and MSc degree from Southeast University, China, in 2006 and 2008, respectively. 

My research interest includes recommender systems, information retrieval and data mining.

Latest News
  • "GB-CENT: Gradient Boosted Categorical Embedding and Numerical Trees" with Qian Zhao, Liangjie Hong, accepted in WWW 2017.
  • "On Sampling Strategies for Neural Network-based Collaborative Filtering" with Ting Chen, Yizhou Sun, Liangjie Hong, accepted in KDD 2017.
  • "A Gradient-based Adaptive Learning Framework for Efficient Personal Recommendation" with Yue Ning, Liangjie Hong, Huzefa Rangwala, and Naren Ramakrishnan accepted in RecSys 2017.
  • "Returning is Believing: Optimizing Long-term User Engagement in Recommender Systems" with Qingyun Wu, Hongning Wang, Liangjie Hong accepted in CIKM 2017.
  • PC member in SIGIR 2018, WWW 2018.