Exploring Machine Learning Privacy/Utility trade-off from a hyperparameters Lens.
Ayoub ArousAmira GuesmiMuhammad Abdullah HanifIhsen AlouaniMuhammad ShafiquePublished in: CoRR (2023)
Keyphrases
- signal to noise ratio
- hyperparameters
- trade off
- machine learning
- model selection
- noise reduction
- noise level
- cross validation
- differential privacy
- closed form
- support vector
- bayesian inference
- random sampling
- bayesian framework
- em algorithm
- gaussian process
- privacy preserving
- sample size
- maximum a posteriori
- gaussian processes
- prior information
- feature selection
- edge detection
- incomplete data
- support vector machine
- incremental learning
- active learning
- computer vision
- data mining
- machine learning methods
- learning algorithm
- parameter space
- machine learning algorithms
- generative model
- markov random field
- data sets
- grid search
- missing values
- supervised learning
- lower bound
- data analysis
- image processing