Information processing method, information processing apparatus, and computer readable storage medium

Patent metadata

Field	Value
Publication number	US-2019279087-A1
Application number	US-201916424108-A
Country	US
Kind code	A1
Filing date	May 28, 2019
Priority date	Jan 13, 2017
Publication date	Sep 12, 2019
Grant date	—

Abstract

An information processing method includes: reading a layer structure and parameters of layers from each of models of two neural networks; and determining a degree of matching between the models of the two neural networks, by comparing layers, of the respective models of the two neural networks, that are configured as a graph-like form in respective hidden layers, in order from an input layer using breadth first search or depth first search, based on similarities between respective layers.

First claim

1 . An information processing method comprising:
reading a layer structure and parameters of layers from each of models of two neural networks; and determining a degree of matching between the models of the two neural networks, by comparing layers, of the respective models of the two neural networks, that are configured as a graph-like form in respective hidden layers, in order from an input layer using breadth first search or depth first search, based on similarities between respective layers.

2 . The information processing method according to claim 1 , wherein the determining the degree of matching between the models of the two neural networks includes, if types of layers to be compared are different, setting a similarity to 0, and not performing comparison in layers in a later stage than the layers subjected to the comparison.

3 . The information processing method according to claim 1 ,
wherein the determining the degree of matching between the models of the two neural networks includes, when convolutional layers are compared, estimating a true filter size with respect to each weight filter of the convolutional layers, modifying parameters of weight filters to be compared to respective estimated true filter sizes, expressing the convolutional layers by vector sets by regarding parameters of each weight filter as one vector, and setting a similarity between the vector sets of the respective convolutional layers to be compared as a similarity between the convolutional layers.

4 . The information processing method according to claim 3 ,
wherein the estimating the true filter size includes, accumulating absolute values of parameters of respective channels in each of the weight filters, and estimating a minimum rectangle that includes all positions at which accumulated values are a predetermined threshold value or more, as a true filter size.

5 . The information processing method according to claim 1 ,
wherein the determining the degree of matching between the models of the two neural networks includes, when full-connected layers are compared, expressing each of the full-connected layers, by regarding weights of each of the full-connected layers as a feature vector, using a vector set, and setting a similarity between the vector sets of the respective full-connected layers to be compared as a similarity between the full-connected layers.

6 . The information processing method according to claim 3 ,
wherein the similarity between the vector sets is obtained by configuring a bipartite graph by obtaining pairs of feature vectors whose distance is a predetermined threshold value or less, and calculating a maximum number of matches by solving a maximum matching problem from the bipartite graph.

7 . The information processing method according to claim 3 ,
wherein the similarity between the vector sets is obtained by quantizing respective feature vectors, and obtaining a similarity between quantization histograms.

8 . The information processing method according to claim 1 ,
wherein the determining the degree of matching between the models of the two neural networks includes, when activation layers are compared, setting, if types of the activation layers are the same, and a distance between parameters is a predetermined threshold value or less, the distance as a similarity, and in other cases, 0 to the similarity.

9 . The information processing method according to claim 8 , wherein the type of the activation layer is a linear connection, a sigmoid function, a hard sigmoid function, a tanh function (hyperbolic tangent function), a softsign function, a softplus function, or a ReLU (Rectified Linear Unit).

10 . The information processing method according to claim 1 ,
wherein the determining the degree of matching between the models of the two neural networks includes, when pooling layers are compared, setting, if types of the pooling layers are the same, and a distance between parameters is a predetermined threshold value or less, the distance as a similarity, and in other cases, 0 to the similarity.

11 . The information processing method according to claim 10 ,
wherein the type of the pooling layer is max pooling or average pooling, and the parameters are a filter size and an interval at which a filter is applied.

12 . An information processing apparatus comprising:
a reading unit configured to read a layer structure and parameters of layers from each of models of two neural networks; and a determining unit configured to determine a degree of matching between the models of the two neural networks, by comparing layers, of the respective models of the two neural networks, that are configured as a graph-like form in respective hidden layers, in order from an input layer using breadth first search or depth first search, based on similarities between respective layers.

13 . A computer readable storage medium storing a program, the program causes, upon being executed by one or more processors of a computer, the computer to execute:
reading a layer structure and parameters of layers from each of models of two neural networks; and determining a degree of matching between the models of the two neural networks, by comparing layers, of the respective models of the two neural networks, that are configured as a graph-like form in respective hidden layers, in order from an input layer using breadth first search or depth first search, based on similarities between respective layers.

Assignees

• Kddi Corp

Inventors

• Uchida Yusuke
• Sakazawa Shigeyuki
• Nagai Yuki

Classifications

• G06N3/084

  Backpropagation, e.g. using gradient descent · CPC title

• G06V30/1988

  Graph matching (graphical image representation G06V30/18181) · CPC title

• G06V30/1916

  Validation; Performance evaluation · CPC title

• G06V30/19173

  Classification techniques · CPC title

• G06N3/045

  Combinations of networks · CPC title