XNORAM: An Efficient Computing-in-Memory Architecture for Binary Convolutional Neural Networks with Flexible Dataflow Mapping.
Shiwei LiuHaozhe ZhuChixiao ChenLihua ZhangC.-J. Richard ShiPublished in: AICAS (2020)
Keyphrases
- convolutional neural networks
- dynamic reconfiguration
- data flow
- design methodology
- database machine
- real time
- convolutional network
- management system
- memory requirements
- neural network
- highly flexible
- memory management
- parallel computing
- associative memory
- limited memory
- memory size
- software architecture
- memory space
- memory efficient
- hardware architecture
- processing elements
- network architecture
- multithreading
- software systems
- multiresolution
- feature extraction