NDPipe: Exploiting Near-data Processing for Scalable Inference and Continuous Training in Photo Storage.
Jungwoo KimSeonggyun OhJaeha KungYeseong KimSungjin LeePublished in: ASPLOS (3) (2024)
Keyphrases
- data processing
- structured prediction
- data analysis
- data storage
- data management
- inference process
- lightweight
- database
- continuous query processing
- training process
- real time
- inference control
- storage systems
- training algorithm
- data acquisition
- online learning
- bayesian networks
- training set
- data sets
- computer systems
- data grids
- databases
- training data
- storage devices
- random access
- highly scalable
- storage requirements
- inference engine
- artificial neural networks
- semi supervised
- supervised learning
- low cost
- test set
- training samples