On the importance of pre-training data volume for compact language models.
Vincent MicheliMartin d'HoffschmidtFrançois FleuretPublished in: EMNLP (1) (2020)
Keyphrases
- language model
- training data
- language modeling
- n gram
- language modelling
- probabilistic model
- query expansion
- speech recognition
- retrieval model
- document retrieval
- learning algorithm
- statistical language models
- context sensitive
- training set
- information retrieval
- decision trees
- test collection
- supervised learning
- ad hoc information retrieval
- vector space model
- smoothing methods
- document ranking
- language models for information retrieval
- document length
- pseudo relevance feedback
- term dependencies
- labeled data
- translation model
- relevance model
- word error rate
- classification accuracy
- information extraction
- language model for information retrieval