CirCNN: Accelerating and Compressing Deep Neural Networks Using Block-Circulant Weight Matrices

@article{Ding2017CirCNNAA,
  title={CirCNN: Accelerating and Compressing Deep Neural Networks Using Block-Circulant Weight Matrices},
  author={Caiwen Ding and Siyu Liao and Yanzhi Wang and Z. Li and N. Liu and Youwei Zhuo and Chao Ching Wang and Xuehai Qian and Y. Bai and Geng Yuan and X. Ma and Yipeng Zhang and J. Tang and Qinru Qiu and X. Lin and Bo Yuan},
  journal={2017 50th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO)},
  year={2017},
  pages={395-408}
}
  • Caiwen Ding, Siyu Liao, +13 authors Bo Yuan
  • Published 2017
  • Computer Science
  • 2017 50th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO)
Large-scale deep neural networks (DNNs) are both compute and memory intensive. [...] Key Method We propose the CirCNN architecture, a universal DNN inference engine that can be implemented in various hardware/software platforms with configurable network architecture (e.g., layer type, size, scales, etc In CirCNN architecture: 1) Due to the recursive property, FFT can be used as the key computing kernel which ensures universal and small-footprint implementations.Expand
148 Citations
PermDNN: Efficient Compressed DNN Architecture with Permuted Diagonal Matrices
  • 49
  • PDF
BlockGNN: Towards Efficient GNN Acceleration Using Block-Circulant Weight Matrices
  • Highly Influenced
  • PDF
ADMM-NN: An Algorithm-Hardware Co-Design Framework of DNNs Using Alternating Direction Methods of Multipliers
  • 57
  • PDF
Block-Circulant Neural Network Accelerator Featuring Fine-Grained Frequency-Domain Quantization and Reconfigurable FFT Modules
  • Highly Influenced
Non-structured DNN Weight Pruning Considered Harmful
  • 6
Structured Weight Matrices-Based Hardware Accelerators in Deep Neural Networks: FPGAs and ASICs
  • 13
  • PDF
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 21 REFERENCES
An Exploration of Parameter Redundancy in Deep Networks with Circulant Projections
  • 230
  • Highly Influential
  • PDF
Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding
  • 4,444
  • Highly Influential
  • PDF
EIE: Efficient Inference Engine on Compressed Deep Neural Network
  • Song Han, Xingyu Liu, +4 authors W. Dally
  • Computer Science
  • 2016 ACM/IEEE 43rd Annual International Symposium on Computer Architecture (ISCA)
  • 2016
  • 1,478
  • Highly Influential
  • PDF
Stripes: Bit-serial deep neural network computing
  • 220
  • Highly Influential
  • PDF
Caffeine: Towards uniformed representation and acceleration for deep convolutional neural networks
  • 272
  • Highly Influential
  • PDF
DaDianNao: A Machine-Learning Supercomputer
  • Yunji Chen, Tao Luo, +8 authors O. Temam
  • Computer Science
  • 2014 47th Annual IEEE/ACM International Symposium on Microarchitecture
  • 2014
  • 925
  • Highly Influential
  • PDF
ESE: Efficient Speech Recognition Engine with Sparse LSTM on FPGA
  • 383
  • Highly Influential
  • PDF
Learning both Weights and Connections for Efficient Neural Network
  • 3,043
  • Highly Influential
  • PDF
Accelerating Binarized Convolutional Neural Networks with Software-Programmable FPGAs
  • 230
  • Highly Influential
  • PDF
Convolutional networks for fast, energy-efficient neuromorphic computing
  • 446
  • Highly Influential
  • PDF
...
1
2
3
...