Boosting Transductive Few-Shot Fine-tuning with Margin-based Uncertainty Weighting and Probability Regularization.
Ran TaoHao ChenMarios SavvidesPublished in: CVPR (2023)
Keyphrases
- fine tuning
- boltzmann distribution
- boosting algorithms
- conditional probabilities
- rademacher complexity
- probability theory
- multi class
- unlabeled data
- early stopping
- viable alternative
- probability measure
- probability distribution
- fine tune
- generalization error
- learning algorithm
- video sequences
- kernel machines
- semi supervised learning
- kernel ridge regression
- feature selection
- fine tuned
- support vector
- semi supervised
- detection algorithm
- machine learning
- gradient boosting
- weighting scheme
- text classification
- ensemble learning
- data dependent
- transductive learning
- labeled data
- support vector machine
- reproducing kernel hilbert space
- active learning
- discriminative learning
- bayesian networks
- cost sensitive
- similarity measure
- video content
- loss function
- video data