On-chip trainable hardware-based deep Q-networks approximating a backpropagation algorithm.
Jangsaeng KimDongseok KwonSung Yun WooWon-Mook KangSoochang LeeSeongbin OhChul-Heung KimJong-Ho BaeByung-Gook ParkJong-Ho LeePublished in: Neural Comput. Appl. (2021)
Keyphrases
- backpropagation algorithm
- low cost
- vlsi implementation
- programmable logic
- back propagation
- single chip
- host computer
- backpropagation learning algorithm
- high bandwidth
- learning rate
- feedforward neural networks
- multilayer feedforward
- neural network
- circuit design
- activation function
- real time
- multi layer perceptron
- ibm zenterprise
- social networks
- multilayer perceptron
- network structure
- hidden layer
- hardware implementation
- chip design
- ibm power processor
- feed forward neural networks
- artificial neural networks
- decision making