Unbiased Learning-to-Rank Needs Unconfounded Propensity Estimation.
Dan LuoLixin ZouQingyao AiZhiyu ChenChenliang LiDawei YinBrian D. DavisonPublished in: SIGIR (2024)
Keyphrases
- learning to rank
- information retrieval
- ranking functions
- loss function
- evaluation measures
- document retrieval
- collaborative filtering
- ranking svm
- direct optimization
- retrieval systems
- query dependent
- directly optimize
- test collection
- evaluation metrics
- text categorization
- data mining
- supervised learning
- semi supervised
- learning to rank algorithms