Method, electronic device, and storage medium for expanding data

Patent metadata

Field	Value
Publication number	US-12165072-B2
Application number	US-202117213952-A
Country	US
Kind code	B2
Filing date	Mar 26, 2021
Priority date	Apr 8, 2020
Publication date	Dec 10, 2024
Grant date	Dec 10, 2024

Abstract

A method, apparatus, device, and storage medium for expanding data are disclosed. The method includes: acquiring a triplet from a knowledge graph; mining a relationship path equivalent to a relationship in the triplet from the knowledge graph, a subject in the triplet being used as a start point of the relationship path, and an object in the triplet being used as an end point of the relationship path; and expanding the triplet based on the relationship path to generate an expanded triplet. This implementation expands the triplet in the knowledge graph, and strengthens the association between the subject and the object in the triplet in a larger context, such that the association between the subject and the object in the triplet is more global.

First claim

What is claimed is:

1. A method for expanding data, comprising:
acquiring a triplet from a knowledge graph;
mining a relationship path equivalent to a relationship in the triplet from the knowledge graph, a subject in the triplet being used as a start point of the relationship path, and an object in the triplet being used as an end point of the relationship path;
expanding the triplet based on the relationship path to generate an expanded triplet;
adding the triplet and the expanded triplet into a training sample set as training samples; and
obtaining, for the training samples in the training sample set, a prediction model by supervised training with a training sample having a missing subject or object as an input, and with the missing subject or object as an output.

2. The method according to claim 1 , wherein the mining the relationship path equivalent to the relationship in the triplet from the knowledge graph comprises:
mining other triplet sequences that statistically cooccur with the subject and the object in the triplet from the knowledge graph; and
sequentially combining relationships in the other triplet sequences to generate the relationship path.

3. The method according to claim 1 , wherein the expanding the triplet based on the relationship path to generate the expanded triplet comprises:
replacing the relationship in the triplet with the relationship path to generate the expanded triplet.

4. The method according to claim 1 , wherein the expanding the triplet based on the relationship path to generate the expanded triplet comprises:
traversing the relationship path with the subject in the triplet as the start point to obtain another object other than the object in the triplet; and
replacing the relationship in the triplet with the relationship path, and replacing the object in the triplet with the another object, to generate the expanded triplet.

5. The method according to claim 1 , wherein the method further comprises:
acquiring a predicted triplet, wherein the predicted triplet has a missing subject or object;
determining a predicted expanded triplet corresponding to the predicted triplet based on the knowledge graph;
inputting the predicted triplet into the prediction model to obtain a confidence degree of a predicted subject or object corresponding to the predicted triplet, and inputting the predicted expanded triplet into the prediction model to obtain a confidence degree of the predicted subject or object corresponding to the predicted expanded triplet; and
computing an average of the confidence degree of the predicted subject or object corresponding to the predicted triplet and the predicted expanded triplet, to determine the missing subject or object of the predicted triplet.

6. An electronic device, comprising:
at least one processor; and
a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, such that the at least one processor can perform operations comprising:
acquiring a triplet from a knowledge graph;
mining a relationship path equivalent to a relationship in the triplet from the knowledge graph, a subject in the triplet being used as a start point of the relationship path, and an object in the triplet being used as an end point of the relationship path;
expanding the triplet based on the relationship path to generate an expanded triplet;
adding the triplet and the expanded triplet into a training sample set as training samples; and
obtaining, for the training samples in the training sample set, a prediction model by supervised training with a training sample having a missing subject or object as an input, and with the missing subject or object as an output.

7. The electronic device according to claim 6 , wherein the mining the relationship path equivalent to the relationship in the triplet from the knowledge graph comprises:
mining other triplet sequences that statistically cooccur with the subject and the object in the triplet from the knowledge graph; and
sequentially combining relationships in the other triplet sequences to generate the relationship path.

8. The electronic device according to claim 6 , wherein the expanding the triplet based on the relationship path to generate the expanded triplet comprises:
replacing the relationship in the triplet with the relationship path to generate the expanded triplet.

9. The electronic device according to claim 6 , wherein the expanding the triplet based on the relationship path to generate the expanded triplet comprises:
traversing the relationship path with the subject in the triplet as the start point to obtain another object other than the object in the triplet; and
replacing the relationship in the triplet with the relationship path, and replacing the object in the triplet with the another object, to generate the expanded triplet.

10. The electronic device according to claim 6 , wherein the operations further comprise:
acquiring a predicted triplet, wherein the predicted triplet has a missing subject or object;
determining a predicted expanded triplet corresponding to the predicted triplet based on the knowledge graph;
inputting the predicted triplet into the prediction model to obtain a confidence degree of a predicted subject or object corresponding to the predicted triplet, and inputting the predicted expanded triplet into the prediction model to obtain a confidence degree of the predicted subject or object corresponding to the predicted expanded triplet; and
computing an average of the confidence degree of the predicted subject or object corresponding to the predicted triplet and the predicted expanded triplet, to determine the missing subject or object of the predicted triplet.

11. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions when executed by a computer, cause the computer to perform operations comprising:
acquiring a triplet from a knowledge graph;
mining a relationship path equivalent to a relationship in the triplet from the knowledge graph, a subject in the triplet being used as a start point of the relationship path, and an object in the triplet being used as an end point of the relationship path;
expanding the triplet based on the relationship path to generate an expanded triplet;
adding the triplet and the expanded triplet into a training sample set as training samples; and
obtaining, for the training samples in the training sample set, a prediction model by supervised training with a training sample having a missing subject or object as an input, and with the missing subject or object as an output.

12. The non-transitory computer-readable storage medium according to claim 11 , wherein the mining the relationship path equivalent to the relationship in the triplet from the knowledge graph comprises:
mining other triplet sequences that statistically cooccur with the subject and the object in the triplet from the knowledge graph; and
sequentially combining relationships in the other triplet sequences to generate the relationship path.

13. The non-transitory computer-readable storage medium according to claim 11 , wherein the expanding the triplet based on the relationship path to generate the expanded triplet comprises:
replacing the relationship in the triplet with the relationship path to generate the expanded triplet.

14. The non-transitory computer-readable storage medium according to claim 11 , wherein the expanding the triplet based on the relationship path to genera

Assignees

• Beijing Baidu Netcom Sci & Tech Co Ltd

Inventors

• Huang Pingping
• Wang Quan
• Jiang Wenbin
• Yuan Pengcheng

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06F18/2148

  characterised by the process organisation or structure, e.g. boosting cascade · CPC title

• G06F18/29

  Graphical models, e.g. Bayesian networks · CPC title

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06F16/288

  Entity relationship models · CPC title