TCRA-LLM: Token Compression Retrieval Augmented Large Language Model for Inference Cost Reduction.
Junyi LiuLiangzhi LiTong XiangBowen WangYiming QianPublished in: CoRR (2023)
Keyphrases
- language model
- cost reduction
- retrieval model
- document retrieval
- query expansion
- test collection
- ad hoc information retrieval
- information retrieval
- language modeling
- language models for information retrieval
- query terms
- cross language retrieval
- n gram
- statistical language models
- document length
- speech recognition
- probabilistic model
- smoothing methods
- document ranking
- query specific
- jelinek mercer
- relevance model
- mixture model
- term dependencies
- language modelling
- pseudo feedback
- web page retrieval
- information retrieval systems
- okapi bm
- cost savings
- vector space model
- pseudo relevance feedback
- context sensitive
- text retrieval
- bayesian networks
- retrieval effectiveness
- translation model
- term frequency
- average precision
- statistical language modeling
- word clouds
- ad hoc retrieval
- retrieval accuracy
- tf idf
- lead time
- vector space
- document collections
- relevance feedback
- state space
- reinforcement learning