On layer-level control of DNN training and its impact on generalization.
Simon CarbonnelleChristophe De VleeschouwerPublished in: CoRR (2018)
Keyphrases
- training process
- lower level
- control system
- multi layer
- supervised learning
- higher level
- cmac neural network
- training error
- control strategies
- optimal control
- control strategy
- test set
- search engine
- robotic systems
- serious games
- training data
- levels of abstraction
- adaptive control
- metadata
- feedforward neural networks
- application layer
- real time