Improving Bi-encoder Document Ranking Models with Two Rankers and Multi-teacher Distillation.
Jaekeol ChoiEuna JungJangwon SuhWonjong RheePublished in: SIGIR (2021)
Keyphrases
- learning to rank
- ranking models
- learning to rank algorithms
- information retrieval
- document retrieval
- retrieval systems
- ranking functions
- normalized discounted cumulative gain
- ranking svm
- loss function
- document collections
- bit rate
- ranking algorithm
- keywords
- benchmark datasets
- directly optimize
- web search
- learning process
- text documents
- evaluation measures
- classification models
- information retrieval systems
- query terms
- data sets
- relevant documents
- machine learning