Found in the Middle: Permutation Self-Consistency Improves Listwise Ranking in Large Language Models.
Raphael TangXinyu ZhangXueguang MaJimmy LinFerhan TurePublished in: CoRR (2023)
Keyphrases
- language model
- learning to rank
- document retrieval
- ranking functions
- information retrieval
- ranking algorithm
- test collection
- language modeling
- learning to rank algorithms
- query dependent
- ranking svm
- retrieval model
- document ranking
- n gram
- probabilistic model
- query expansion
- evaluation measures
- ranking models
- text retrieval
- statistical language models
- loss function
- evaluation metrics
- ad hoc information retrieval
- query terms
- retrieval systems
- pseudo relevance feedback
- relevance judgments
- document length
- collaborative filtering
- pairwise
- vector space model
- user feedback
- retrieval effectiveness
- web search engines
- relevance model
- smoothing methods
- balancing exploration and exploitation
- expert search
- query specific