Binarized neural networks for resource-efficient hashing with minimizing quantization loss

Zheng, Feng1; Deng, Cheng2; Huang, Heng3,4

2019

1045-0823

28th International Joint Conference on Artificial Intelligence, IJCAI 2019

2019-August

1032-1040

Macao, China

International Joint Conferences on Artificial Intelligence

In order to solve the problem of memory consumption and computational requirements, this paper proposes a novel learning binary neural network framework to achieve a resource-efficient deep hashing. In contrast to floating-point (32-bit) full-precision networks, the proposed method achieves a 32x model compression rate. At the same time, computational burden in convolution is greatly reduced due to efficient Boolean operations. To this end, in our framework, a new quantization loss defined between the binary weights and the learned real values is minimized to reduce the model distortion, while, by minimizing a binary entropy function, the discrete optimization is successfully avoided and the stochastic gradient descend method can be used smoothly. More importantly, we provide two theories to demonstrate the necessity and effectiveness of minimizing the quantization losses for both weights and activations. Numerous experiments show that the proposed method can achieve fast code generation without sacrificing accuracy.
© 2019 International Joint Conferences on Artificial Intelligence. All rights reserved.

第一

EI

U.S. Navy[DBI 1836866] ; U.S. Navy[IIS 1836938] ; U.S. Navy[IIS 1836945] ; U.S. Navy[IIS 1837956] ; U.S. Navy[IIS 1838627] ; U.S. Navy[IIS 1845666] ; U.S. Navy[IIS 1852606]

20194607696707

EV Compendex

1.Department of Computer Science and Engineering, Southern University of Science and Technology, China
2.School of Electronic Enigineering, Xidian University, China
3.Department of Electrical and Computer Engineering, University of Pittsburgh, United States
4.JD Finance America Corporation, China

Zheng, Feng,Deng, Cheng,Huang, Heng. Binarized neural networks for resource-efficient hashing with minimizing quantization loss[C]:International Joint Conferences on Artificial Intelligence,2019:1032-1040.