Calibration and regret bounds for order-preserving surrogate losses in learning to rank.
Clément CalauzènesNicolas UsunierPatrick GallinariPublished in: Mach. Learn. (2013)
Keyphrases
- learning to rank
- order preserving
- regret bounds
- online learning
- lower bound
- information retrieval
- ranking functions
- linear regression
- loss function
- upper bound
- document retrieval
- hash functions
- bregman divergences
- retrieval systems
- search engine
- supervised learning
- document collections
- least squares
- optimal solution
- face recognition
- feature selection