DRAGON: A Dynamic Scheduling and Scaling Controller for Managing Distributed Deep Learning Jobs in Kubernetes Cluster.
Chan-Yi LinTing-An YehJerry ChouPublished in: CLOSER (2019)
Keyphrases
- dynamic scheduling
- deep learning
- flexible manufacturing systems
- unsupervised learning
- scheduling policies
- unsupervised feature learning
- machine learning
- process planning
- control system
- grid services
- manufacturing systems
- weakly supervised
- mental models
- clustering algorithm
- neural network
- loosely coupled
- data points
- data mining
- scheduling algorithm
- distributed environment
- real time control
- segmentation algorithm
- peer to peer