Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
COMPOSITE BINARY DECOMPOSITION NETWORK
Document Type and Number:
WIPO Patent Application WO/2020/061884
Kind Code:
A1
Abstract:
Embodiments are directed to a composite binary decomposition network. An embodiment of a computer-readable storage medium includes executable computer program instructions for transforming a pre-trained first neural network into a binary neural network by processing layers of the first neural network in a composite binary decomposition process, where the first neural network having floating point values representing weights of various layers of the first neural network. The composite binary decomposition process includes a composite operation to expand real matrices or tensors into a plurality of binary matrices or tensors, and a decompose operation to decompose one or more binary matrices or tensors of the plurality of binary matrices or tensors into multiple lower rank binary matrices or tensors.

Inventors:
LI JIANGUO (CN)
CHEN YURONG (CN)
WANG ZHENG (CN)
Application Number:
PCT/CN2018/107886
Publication Date:
April 02, 2020
Filing Date:
September 27, 2018
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
INTEL CORP (US)
LI JIANGUO (CN)
CHEN YURONG (CN)
International Classes:
G06N3/06
Foreign References:
US10032110B22018-07-24
US20170286830A12017-10-05
CN108334945A2018-07-27
CN106816147A2017-06-09
Other References:
HYEONUK, KIM ET AL.: "A Kernel Decomposition Architecture for Binary-weight Convolutional Neural Networks", 2017 54TH ACM/EDAC/ IEEE DESIGN AUTOMATION CONFERENCE, 22 June 2017 (2017-06-22), XP058367856
LI ZEFAN ET AL.: "2017 IEEE International Conference on Computer Vision (ICCV)", 22 October 2017, IEEE, article "Performance guaranteed network acceleration via high-order residual quantization", pages: 2603 - 2611
See also references of EP 3857460A4
Attorney, Agent or Firm:
SHANGHAI PATENT & TRADEMARK LAW OFFICE, LLC (CN)
Download PDF: