Preserving Pre-trained Features Helps Calibrate Fine-tuned Language Models.
Guande HeJianfei ChenJun ZhuPublished in: ICLR (2023)
Keyphrases
- language model
- language modeling
- pre trained
- document retrieval
- fine tuned
- information retrieval
- n gram
- retrieval model
- query expansion
- feature extraction
- probabilistic model
- image features
- language modelling
- speech recognition
- document ranking
- test collection
- feature vectors
- feature space
- statistical language models
- training examples
- relevance model
- general purpose
- text mining
- high dimensional
- computer vision