Nebula-I: A General Framework for Collaboratively Training Deep Learning Models on Low-Bandwidth Cloud Clusters.
Yang XiangZhihua WuWeibao GongSiyu DingXianjie MoYuang LiuShuohuan WangPeng LiuYongshuai HouLong LiBin WangShaohuai ShiYaqian HanYue YuGe LiYu SunYanjun MaDianhai YuPublished in: CoRR (2022)
Keyphrases
- learning models
- low bandwidth
- wireless networks
- machine learning
- loss function
- mobile devices
- learning algorithm
- high bandwidth
- classification models
- mobile computing
- video conferencing
- semi supervised learning
- machine learning algorithms
- communication cost
- learning tasks
- learning problems
- conditional random fields
- wireless communication
- database server
- training set
- supervised learning
- computing environments
- mobile networks
- training data
- mobile environments
- cloud computing
- multi label
- data sets
- computer simulation
- data points
- active learning
- pairwise
- support vector