General Framework for Self-Supervised Model Priming for Parameter-Efficient Fine-tuning.
Shih-Cheng HuangShih-Heng WangMin-Han ShihSaurav SahayHung-yi LeePublished in: CoRR (2022)
Keyphrases
- fine tuning
- formal model
- modeling framework
- theoretical framework
- computational model
- conceptual framework
- probabilistic model
- lightweight
- theoretical foundation
- parameter values
- modeling method
- mathematical model
- experimental data
- general purpose
- generic model
- decision theoretic
- unified model
- decision model
- data sets
- bayesian framework
- conceptual model
- closed form
- theoretical analysis
- model selection
- input data
- management system
- cost function
- special case
- prior knowledge