Copycat CNN: Are Random Non-Labeled Data Enough to Steal Knowledge from Black-box Models?
Jacson Rodrigues Correia da SilvaRodrigo Ferreira BerrielClaudine BadueAlberto F. De SouzaThiago Oliveira-SantosPublished in: CoRR (2021)
Keyphrases
- black box
- labeled data
- prior knowledge
- unlabeled data
- semi supervised
- semi supervised learning
- active learning
- text classification
- supervised learning
- training data
- supervised methods
- domain adaptation
- transfer learning
- knowledge management
- black boxes
- data points
- background knowledge
- white box
- database
- semi supervised classification
- transferring knowledge
- supervised learning algorithms
- labeled examples
- white box testing
- co training
- class labels
- text mining
- probabilistic model
- knowledge transfer
- label propagation
- learning models
- labeled training data
- text categorization
- integration testing
- knowledge representation
- learning algorithm