ML-Stealer: Stealing Prediction Functionality of Machine Learning Models with Mere Black-Box Access.
Gaoyang LiuShijie WangBorui WanZekun WangChen WangPublished in: TrustCom (2021)
Keyphrases
- black box
- machine learning models
- predictive model
- spam filtering
- black boxes
- machine learning algorithms
- white box
- machine learning approaches
- prediction accuracy
- maximum likelihood
- integration testing
- prediction model
- learning models
- historical data
- test cases
- white box testing
- machine learning
- learning algorithm
- data mining methods
- classification rules
- software engineering
- probabilistic model
- hidden markov models
- feature space
- database systems