DP-SGD Without Clipping: The Lipschitz Neural Network Way.
Louis BéthuneThomas MassenaThibaut BoissinAurélien BelletFranck MamaletYannick PrudentCorentin FriedrichMathieu SerrurierDavid VigourouxPublished in: ICLR (2024)
Keyphrases
- neural network
- dynamic programming
- neural network model
- artificial neural networks
- back propagation
- pattern recognition
- pointwise
- fuzzy logic
- prediction model
- genetic algorithm
- neural nets
- metric space
- self organizing maps
- fuzzy artmap
- backpropagation neural network
- neural network is trained
- image reconstruction from projections
- stochastic gradient descent
- learning vector quantization
- volume rendering
- feed forward neural networks
- data sets
- fuzzy neural network
- training algorithm
- hidden layer
- multilayer perceptron
- image restoration
- model selection
- online learning
- learning algorithm