XOMA: exclusive on-chip memory architecture for energy-efficient deep learning acceleration.
Hyeon Uk SimJason Helge AndersonJongeun LeePublished in: ASP-DAC (2019)
Keyphrases
- energy efficient
- deep learning
- wireless sensor networks
- level parallelism
- multi core architecture
- energy consumption
- multithreading
- memory access
- sensor networks
- unsupervised learning
- machine learning
- base station
- high speed
- energy efficiency
- data transmission
- mental models
- unsupervised feature learning
- management system
- memory management
- low cost
- associative memory
- parallel computing
- object recognition
- weakly supervised
- real time
- processing elements
- routing protocol
- cloud computing
- sensor nodes