Amplifying Training Data Exposure through Fine-Tuning with Pseudo-Labeled Memberships.
Myung Gyo OhHong Eun AhnLeo Hyun ParkTaekyoung KwonPublished in: CoRR (2024)
Keyphrases
- fine tuning
- training data
- training set
- supervised learning
- fine tune
- test data
- data sets
- learning algorithm
- viable alternative
- labeled training data
- test set
- training examples
- classification accuracy
- decision trees
- classification models
- support vector machine
- training instances
- training dataset
- fine tuned
- prior knowledge
- unlabeled data
- naive bayes
- class labels
- negative examples
- active learning
- training process
- training and test data
- pairwise
- data points