# US 2026/0093773 A1 · Information processing apparatus, information processing method and non-transitory computer-readable storage medium

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0093773 A1
- Document ID: US20260093773A1
- Status: Published application
- Kind: A1
- Application number: 19/335,086
- Filed: 2025-09-22
- Published: 2026-04-02
- Earliest claimed: 2024-09-30
- Assignees: Canon Kabushiki Kaisha
- Applicants: Canon Kabushiki Kaisha
- Inventors: Shigeo KODAMA
- Web page: https://patentgrep.com/patent/US20260093773A1

## CPC

- G06F 17/15: Correlation function computation {including computation of convolution operations (arithmetic circuits for sum of products per se, e.g. multiply-accumulators G06F7/5443; digital filters, e.g. FIR, IIR, adaptive filters H03H17/00)}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/048: Activation functions
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An information processing apparatus for performing a graph convolution operation on a graph comprises a first acquisition unit configured to acquire from a list comprising weights of nodes of the graph and connection destinations of those nodes and for which information of the portion, in an adjacency matrix representing connection relationships of nodes in the graph, where values representing the connection relationships are non-zero has been extracted, a weight and a connection destination of a node of the graph, a second acquisition unit configured to acquire, based on the connection destination, data to be inputted into a computing unit from among computation target data, and a computation unit configured to use the computing unit to perform an operation using the data acquired by the second acquisition unit and the weight of node acquired by the first acquisition unit.

## Claims

### Claim 1

An information processing apparatus for performing a graph convolution operation on a graph, the apparatus comprising:
a first acquisition unit configured to acquire from a list comprising weights of nodes of the graph and connection destinations of those nodes and for which information of the portion, in an adjacency matrix representing connection relationships of nodes in the graph, where values representing the connection relationships are non-zero has been extracted, a weight and a connection destination of a node of the graph;
a second acquisition unit configured to acquire, based on the connection destination, data to be inputted into a computing unit from among computation target data; and
a computation unit configured to use the computing unit to perform an operation using the data acquired by the second acquisition unit and the weight of node acquired by the first acquisition unit.

### Claim 2 (depends on 1)

The information processing apparatus according to claim 1, wherein the second acquisition unit, based on the connection destination, generates address information of the data to be inputted into the computing unit from among the computation target data, and acquires the data to be inputted into the computing unit based on that address information.

### Claim 3 (depends on 1)

The information processing apparatus according to claim 1, wherein the computation target data is a result of a multiply-accumulate operation on feature data of nodes of the graph and trained parameters.

### Claim 4 (depends on 3)

The information processing apparatus according to claim 3, wherein
in a case where the number of channels of the feature data is more than the number of channels of a computation result of the computation unit, the computation target data is a result of a multiply-accumulate operation on feature data of nodes of the graph and trained parameters, and
in a case where the number of channels of the feature data is less than the number of channels of the computation result of the computation unit, the computation target data is feature data of nodes of the graph.

### Claim 5 (depends on 1)

The information processing apparatus according to claim 1, wherein the computation unit applies a Relu function to a result of a multiply-accumulate operation on the data acquired by the second acquisition unit and the weight of the node acquired by the first acquisition unit.

### Claim 6 (depends on 5)

The information processing apparatus according to claim 5, wherein the computation unit applies the Relu function to a result of a multiply-accumulate operation for a first of two graph convolution layers.

### Claim 7 (depends on 1)

The information processing apparatus according to claim 1, further comprising a classification unit configured to, based on a result of the operation by the computation unit, perform a category classification on the graph.

### Claim 8 (depends on 7)

The information processing apparatus according to claim 7, wherein the classification unit performs the category classification based on a result of the operation by the computation unit for the second of two graph convolution layers.

### Claim 9 (depends on 1)

The information processing apparatus according to claim 1, further comprising a duplication unit configured to hold a multiply-accumulate result by the computation unit and repeatedly output the same result.

### Claim 10 (depends on 1)

The information processing apparatus according to claim 1, wherein the computation unit performs parallel processing on a plurality of feature vectors in different spaces of a feature map.

### Claim 11 (depends on 1)

The information processing apparatus according to claim 1, wherein the first acquisition unit acquires the list, which is structured to list elements in which a weight of a node in the graph indicating an influence between nodes and a connection destination of data to which that weight from a connection source node to a connection destination node is applied in a convolution form a pair, having excluded elements for which there is no connection between nodes in the graph structure.

### Claim 12 (depends on 11)

The information processing apparatus according to claim 11, wherein the first acquisition unit reads the list from consecutive regions in memory.

### Claim 13 (depends on 1)

The information processing apparatus according to claim 1, wherein the computation unit performs a convolution operation using the data acquired by the second acquisition unit and the weight of the node acquired by the first acquisition unit, and outputs a non-zero graph weight and a connection destination of data corresponding to that weight.

### Claim 14 (depends on 13)

The information processing apparatus according to claim 13, wherein the computation unit writes, in consecutive regions in memory, elements in which a weight of a node of the graph and a connection destination of data corresponding to that weight form a pair.

### Claim 15

An information processing method performed by an information processing apparatus for performing a graph convolution operation on a graph, the method comprising:
acquiring from a list comprising weights of nodes of the graph and connection destinations of those nodes and for which information of the portion, in an adjacency matrix representing connection relationships of nodes in the graph, where values representing the connection relationships are non-zero has been extracted, a weight and a connection destination of a node of the graph;
acquiring, based on the connection destination, data to be inputted into a computing unit from among computation target data; and
using the computing unit to perform an operation using the acquired data and the acquired weight of node.

### Claim 16

A non-transitory computer-readable storage medium storing a computer program for causing a computer operable to perform a graph convolution operation on a graph to function as:
a first acquisition unit configured to acquire from a list comprising weights of nodes of the graph and connection destinations of those nodes and for which information of the portion, in an adjacency matrix representing connection relationships of nodes in the graph, where values representing the connection relationships are non-zero has been extracted, a weight and a connection destination of a node of the graph;
a second acquisition unit configured to acquire, based on the connection destination, data to be inputted into a computing unit from among computation target data; and
a computation unit configured to use the computing unit to perform an operation using the data acquired by the second acquisition unit and the weight of node acquired by the first acquisition unit.
