Analog/Mixed-Signal Hardware Error Modeling for Deep Learning Inference.
Angad S. RekhiBrian ZimmerNikola NedovicNingxi LiuRangharajan VenkatesanMiaorong WangBrucek KhailanyWilliam J. DallyC. Thomas GrayPublished in: DAC (2019)
Keyphrases
- high resolution
- mixed signal
- deep learning
- low power
- multi channel
- digital circuits
- low cost
- restricted boltzmann machine
- image processing
- cmos technology
- machine learning
- unsupervised learning
- high speed
- unsupervised feature learning
- power consumption
- hardware and software
- weakly supervised
- mental models
- bayesian networks
- digital signal processing
- real time
- analog to digital converter
- hardware implementation
- data flow
- circuit design
- input image
- pattern recognition