AdapterBias: Parameter-efficient Token-dependent Representation Shift for Adapters in NLP Tasks.
Chin-Lun FuZih-Ching ChenYun-Ru LeeHung-yi LeePublished in: CoRR (2022)
Keyphrases
- natural language
- input output
- natural language processing
- text mining
- computationally expensive
- image representation
- field of natural language processing
- neural network
- compact representations
- input parameters
- feature representation
- parameter values
- machine translation
- named entities
- query processing
- genetic algorithm
- machine learning
- data mining