Improving Deep Learning with a customizable GPU-like FPGA-based accelerator.
Mirko GagliardiEdoardo FusellaAlessandro CilardoPublished in: PRIME (2018)
Keyphrases
- deep learning
- parallel implementation
- unsupervised feature learning
- unsupervised learning
- field programmable gate array
- weakly supervised
- machine learning
- restricted boltzmann machine
- mental models
- deep architectures
- parallel computing
- hardware implementation
- pattern recognition
- image processing
- general purpose
- probabilistic model
- pairwise