DeepX: A Software Accelerator for Low-Power Deep Learning Inference on Mobile Devices.
Nicholas D. LaneSourav BhattacharyaPetko GeorgievClaudio ForlivesiLei JiaoLorena QendroFahim KawsarPublished in: IPSN (2016)
Keyphrases
- low power
- deep learning
- mobile devices
- power consumption
- low cost
- high speed
- single chip
- unsupervised learning
- machine learning
- unsupervised feature learning
- logic circuits
- mental models
- low power consumption
- embedded systems
- bayesian networks
- image sensor
- cmos technology
- weakly supervised
- real time
- object recognition
- multi channel
- field programmable gate array
- image processing
- feature selection
- power dissipation
- mixed signal