Enhancing cross-domain transferability of black-box adversarial attacks on speaker recognition systems using linearized backpropagation.
Umang PatelShruti BhilareAvik HatiPublished in: Pattern Anal. Appl. (2024)
Keyphrases
- black box
- back propagation
- cross domain
- black boxes
- speaker recognition
- neural network
- artificial neural networks
- neural nets
- feed forward
- transfer learning
- multilayer perceptron
- nearest neighbor
- data sets
- knowledge transfer
- gaussian mixture model
- decision trees
- bp neural network
- hidden layer
- text classification
- learning algorithm
- feed forward neural networks
- machine learning
- active learning