Title:
METHODS AND SYSTEMS USING IMPROVED TRAINING AND LEARNING FOR DEEP NEURAL NETWORKS
Document Type and Number:
WIPO Patent Application WO/2018/184222
Kind Code:
A1
Abstract:
Methods and systems are disclosed using improved training and learning for deep neural networks. In one example, a deep neural network includes a plurality of layers, and each layer has a plurality of nodes. For each L layer in the plurality of layers, the nodes of each L layer are randomly connected to nodes in a L+1 layer. For each L+1 layer in the plurality of layers, the nodes of each L+1 layer are connected to nodes in a subsequent L layer in a one-to-one manner. Parameters related to the nodes of each L layer are fixed. Parameters related to the nodes of each L+1 layers are updated, and L is an integer starting with 1. In another example, a deep neural network includes an input layer, output layer, and a plurality of hidden layers. Inputs for the input layer and labels for the output layer are determined related to a first sample. Similarity between different pairs of inputs and labels between a second sample with the first sample is estimated using Gaussian regression process.
Inventors:
GUO YIWEN (CN)
YAO ANBANG (CN)
CAI DONGQI (CN)
WANG LIBIN (CN)
XU LIN (CN)
HU PING (CN)
WANG SHANDONG (CN)
CHENG WENHUA (CN)
CHEN YURONG (CN)
YAO ANBANG (CN)
CAI DONGQI (CN)
WANG LIBIN (CN)
XU LIN (CN)
HU PING (CN)
WANG SHANDONG (CN)
CHENG WENHUA (CN)
CHEN YURONG (CN)
Application Number:
PCT/CN2017/079771
Publication Date:
October 11, 2018
Filing Date:
April 07, 2017
Export Citation:
Assignee:
INTEL CORP (US)
GUO YIWEN (CN)
YAO ANBANG (CN)
CAI DONGQI (CN)
WANG LIBIN (CN)
XU LIN (CN)
HU PING (CN)
WANG SHANDONG (CN)
CHENG WENHUA (CN)
CHEN YURONG (CN)
GUO YIWEN (CN)
YAO ANBANG (CN)
CAI DONGQI (CN)
WANG LIBIN (CN)
XU LIN (CN)
HU PING (CN)
WANG SHANDONG (CN)
CHENG WENHUA (CN)
CHEN YURONG (CN)
International Classes:
G06N3/02; G06N3/04; G06N3/08
Foreign References:
| US20160335536A1 | 2016-11-17 | |||
| CN106355246A | 2017-01-25 | |||
| CN102859538A | 2013-01-02 |
Other References:
SHIJIE ET AL.: "Heterogeneous blocked CPU-GPU accelerate scheme for large scale extreme learning machine", NEUROCOMPUTING, vol. 261, 25 October 2017 (2017-10-25), pages 153 - 163, XP055738992, DOI: 10.1016/j.neucom.2016.05.112
WANG ET AL.: "An efficient and effective convolutional auto-encoder extremelearning machine networkfor 3d feature learning", NEUROCOMPUTING, vol. 174, 22 January 2016 (2016-01-22), pages 988 - 998
See also references of EP 3607495A4
WANG ET AL.: "An efficient and effective convolutional auto-encoder extremelearning machine networkfor 3d feature learning", NEUROCOMPUTING, vol. 174, 22 January 2016 (2016-01-22), pages 988 - 998
See also references of EP 3607495A4
Attorney, Agent or Firm:
NTD PATENT AND TRADEMARK AGENCY LIMITED (CN)
Download PDF:
Previous Patent: METHOD AND APPARATUS FOR USE IN REMOTE DIAGNOSIS OF VEHICLE
Next Patent: METHOD AND DEVICE FOR VIDEO COMPRESSION PROCESSING
Next Patent: METHOD AND DEVICE FOR VIDEO COMPRESSION PROCESSING