Breaking the computation and communication abstraction barrier in distributed machine learning workloads.
Abhinav JangdaJun HuangGuodong LiuAmir Hossein Nodehi SabetSaeed MalekiYoushan MiaoMadanlal MusuvathiTodd MytkowiczOlli SaarikiviPublished in: ASPLOS (2022)
Keyphrases
- machine learning
- communication overhead
- concurrent processes
- distributed control
- communication cost
- pattern recognition
- spatially distributed
- information dissemination
- multi party
- fully distributed
- computer networks
- multi agent
- cooperative
- distributed computation
- group communication
- multimedia communication
- open systems
- explanation based learning
- computer systems
- distributed environment
- computer vision
- distributed network
- data mining
- feature selection
- high level
- database systems
- distributed systems
- information sharing
- artificial intelligence
- learning systems
- decision trees
- communication networks
- reinforcement learning
- data analysis
- information extraction
- learning algorithm
- single point of failure
- active learning
- knowledge representation
- exchange information
- text classification
- communication systems
- lightweight
- inductive logic programming
- machine learning algorithms
- learning tasks
- statistical methods
- machine learning methods