Universal approximation property of invertible neural networks.
Isao IshikawaTakeshi TeshimaKoichi TojoKenta OonoMasahiro IkedaMasashi SugiyamaPublished in: CoRR (2022)
Keyphrases
- neural network
- pattern recognition
- genetic algorithm
- fuzzy logic
- relative error
- artificial neural networks
- closed form
- back propagation
- self organizing maps
- approximation error
- fuzzy systems
- error bounds
- neural network model
- efficient computation
- error tolerance
- fault diagnosis
- fuzzy sets
- evolutionary algorithm
- multilayer perceptron
- desirable properties
- decision trees
- training algorithm
- decision making
- learning algorithm
- approximation methods