CAPES: unsupervised storage performance tuning using neural network-based deep reinforcement learning.
Yan LiKenneth ChangOceane BelEthan L. MillerDarrell D. E. LongPublished in: SC (2017)
Keyphrases
- reinforcement learning
- unsupervised learning
- supervised learning
- deep architectures
- data storage
- function approximation
- storage and retrieval
- neural network
- semi supervised
- transfer learning
- markov decision processes
- learning algorithm
- temporal difference learning
- file system
- multi agent
- deep learning
- temporal difference
- unsupervised manner
- multi agent reinforcement learning
- reinforcement learning algorithms
- machine learning
- optimal control
- dynamic programming
- state space
- storage space
- optimal policy
- random access
- robotic control
- active learning
- policy search
- storage systems
- data driven
- database administrators
- learning tasks
- weakly supervised
- learning problems
- storage requirements
- supervised classification