Bounding training data reconstruction in DP-SGD.
Jamie HayesBorja BalleSaeed MahloujifarPublished in: NeurIPS (2023)
Keyphrases
- training data
- dynamic programming
- upper bound
- learning algorithm
- test data
- decision trees
- data sets
- stochastic gradient descent
- test set
- discrete tomography
- three dimensional
- image reconstruction
- classification accuracy
- reconstruction method
- labeled data
- supervised learning
- training examples
- training instances
- training process
- compressed sensing
- reconstruction process
- tomographic reconstruction
- learned from training data
- unlabeled data
- reconstruction error
- matrix factorization
- semi supervised learning
- support vector machine
- active learning
- high resolution
- machine learning
- neural network