RL-DARTS: Differentiable neural architecture search via reinforcement-learning-based meta-optimizer.
Dong PangXinyi LeXinping GuanPublished in: Knowl. Based Syst. (2021)
Keyphrases
- reinforcement learning
- neural architecture
- search algorithm
- reinforcement learning algorithms
- function approximation
- state space
- model free
- markov decision processes
- partially observable domains
- learning algorithm
- feed forward
- search space
- neural network
- optimal policy
- activation function
- action selection
- action space
- machine learning
- transfer learning
- learning classifier systems
- training data
- temporal difference
- rl algorithms
- autonomous learning