Framing RNN as a kernel method: A neural ODE approach.
Adeline FermanianPierre MarionJean-Philippe VertGérard BiauPublished in: NeurIPS (2021)
Keyphrases
- kernel methods
- recurrent neural networks
- neural model
- nearest neighbor
- network architecture
- kernel function
- learning problems
- neural network
- support vector
- feature space
- machine learning
- support vector machine
- reproducing kernel
- learning tasks
- kernel matrix
- ordinary differential equations
- feed forward
- artificial neural networks