Found in the Middle: Permutation Self-Consistency Improves Listwise Ranking in Large Language Models.
Raphael TangXinyu Crystina ZhangXueguang MaJimmy LinFerhan TurePublished in: NAACL-HLT (2024)
Keyphrases
- language model
- learning to rank
- document retrieval
- ranking functions
- information retrieval
- ranking algorithm
- test collection
- language modeling
- learning to rank algorithms
- query dependent
- retrieval model
- n gram
- ranking svm
- probabilistic model
- document ranking
- evaluation measures
- ranking models
- query expansion
- loss function
- statistical language models
- pseudo relevance feedback
- query terms
- web search
- ad hoc information retrieval
- pairwise
- relevance model
- document length
- smoothing methods
- evaluation metrics
- user feedback
- benchmark datasets
- vector space model
- collaborative filtering
- machine learning
- retrieval effectiveness
- information retrieval systems
- web search engines
- classification accuracy
- language models for information retrieval
- query specific
- xml retrieval
- text retrieval
- decision trees
- relevant documents
- retrieval systems
- supervised learning