A linearized framework and a new benchmark for model selection for fine-tuning.
Aditya DeshpandeAlessandro AchilleAvinash RavichandranHao LiLuca ZancatoCharless C. FowlkesRahul BhotikaStefano SoattoPietro PeronaPublished in: CoRR (2021)
Keyphrases
- model selection
- fine tuning
- cross validation
- statistical inference
- mixture model
- parameter estimation
- hyperparameters
- statistical learning
- machine learning
- error estimation
- variable selection
- motion segmentation
- regression model
- bayesian learning
- bayesian information criterion
- model selection criteria
- automatic model selection
- selection criterion
- meta learning
- generalization error
- sample size
- gaussian process
- viable alternative
- probabilistic model
- feature selection