OCEAN: An on-chip incremental-learning enhanced processor with gated recurrent neural network accelerators.
Chixiao ChenHongwei DingHuwan PengHaozhe ZhuRui MaPeiyong ZhangXiaolang YanYu WangMingyu WangHao MinC.-J. Richard ShiPublished in: ESSCIRC (2017)
Keyphrases
- incremental learning
- recurrent neural networks
- single chip
- ibm power processor
- low power
- complex valued
- low cost
- neural network
- feed forward
- high speed
- echo state networks
- hidden layer
- instruction set
- fuzzy artmap
- recurrent networks
- clock frequency
- error resilience
- reservoir computing
- power management
- incremental learning algorithm
- learning process
- supervised learning
- feedforward neural networks
- processor core
- image sensor
- neuro fuzzy
- artificial neural networks
- batch learning
- semi supervised
- chip design
- batch mode
- floating point
- neural network structure
- input output
- pairwise
- machine learning