CMR Scaling Law: Predicting Critical Mixture Ratios for Continual Pre-training of Language Models.
Jiawei GuZacc YangChuanghao DingRui ZhaoFei TanPublished in: CoRR (2024)
Keyphrases
- language model
- mixture model
- language modeling
- probabilistic model
- n gram
- retrieval model
- query expansion
- speech recognition
- language modelling
- information retrieval
- document retrieval
- smoothing methods
- ad hoc information retrieval
- expectation maximization
- statistical language models
- test collection
- training set
- query terms
- relevance model
- document ranking
- context sensitive
- vector space model
- text retrieval
- language model for information retrieval
- term dependencies
- query specific
- language models for information retrieval
- ad hoc retrieval
- translation model
- passage retrieval
- retrieval effectiveness
- error rate
- generative model
- em algorithm
- information extraction
- decision trees