ED2LM: Encoder-Decoder to Language Model for Faster Document Re-ranking Inference.
Kai HuiHonglei ZhuangTao ChenZhen QinJing LuDara BahriJi MaJai Prakash GuptaCícero Nogueira dos SantosYi TayDonald MetzlerPublished in: ACL (Findings) (2022)
Keyphrases
- language model
- document ranking
- document length
- ad hoc information retrieval
- language modeling
- document retrieval
- video codec
- query specific
- distributed video coding
- pseudo feedback
- probabilistic model
- query expansion
- n gram
- language modelling
- information retrieval
- smoothing methods
- ranking functions
- document representation
- retrieval model
- document level
- inter document similarities
- human relevance judgments
- expert search
- relevance model
- query terms
- vector space model
- mixture model
- bayesian networks
- word clouds
- test collection
- retrieved documents
- context sensitive
- bit rate
- language model for information retrieval
- pseudo relevance feedback
- probabilistic retrieval models
- term dependencies
- translation model
- language modeling framework
- tf idf
- web search