Trajectory-dependent Generalization Bounds for Deep Neural Networks via Fractional Brownian Motion.
Chengli TanJiangshe ZhangJunmin LiuPublished in: CoRR (2022)
Keyphrases
- generalization bounds
- fractional brownian motion
- neural network
- long range
- non stationary
- learning theory
- generalization ability
- data dependent
- ranking algorithm
- statistical learning theory
- linear classifiers
- vc dimension
- model selection
- learning problems
- fractal dimension
- artificial neural networks
- financial markets
- bp neural network
- random fields
- pattern recognition
- fault diagnosis
- learning machines
- neural network model
- ranking functions
- statistical learning
- back propagation
- kernel machines
- information retrieval
- conditional random fields
- training algorithm
- multiresolution
- reinforcement learning
- data mining