Pre-trained Language Models Can be Fully Zero-Shot Learners.
Xuandong ZhaoSiqi OuyangZhiguo YuMing WuLei LiPublished in: ACL (1) (2023)
Keyphrases
- language model
- pre trained
- language modeling
- training data
- document retrieval
- n gram
- information retrieval
- language modelling
- probabilistic model
- speech recognition
- query expansion
- e learning
- training examples
- test collection
- statistical language models
- control signals
- retrieval model
- learning process
- learning environment
- relevance model
- document ranking
- smoothing methods
- object recognition
- semi supervised
- feature space
- computer vision