IP Library Granted Patent US 12,373,653
Granted Patent B2
US 12,373,653 · App. 17/821,244 · Granted Jul 29, 2025

Text checking method based on knowledge graph, electronic device, and medium

Inventors: Tianxing Yang (Beijing, CN); Chen Yang (Beijing, CN); Zhuo Sun (Beijing, CN)
Assignee: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.
G06F40/40G06N5/02
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,373,653
App. No.
17/821,244
Granted
Jul 29, 2025
Kind
B2
Abstract

The present disclosure provides a text checking method based on a knowledge graph, an electronic device and a medium. The method includes: obtaining a text of a contract to be checked; establishing a knowledge graph according to the text of the contract to be checked, the knowledge graph including a plurality of attributes and a plurality of attribute values corresponding to the plurality of attributes; and checking the text of the contract to be checked based on the knowledge graph.

Claims (64)

1. A text checking method based on a knowledge graph, performed by a text checking apparatus based on the knowledge graph, comprising:

obtaining a text of a contract to be checked;

establishing a knowledge graph according to the text of the contract to be checked, wherein the knowledge graph comprises a plurality of attributes and a plurality of attribute values corresponding to the plurality of attributes;

checking the text of the contract to be checked based on the knowledge graph,

generating a target check result based on a result in at least one dimension; and

determining whether the text of the contract needs to be modified based on the target check result, comprising:

in the case that a result of one dimension is failed, determining that the target check result is failed and the contract text is modified; and in the case that all the results in the at least one dimension are successful, determining that the target check result is successful and the contract text does not need to be modified,

wherein checking the text of the contract to be checked based on the knowledge graph comprises:

checking the text of the contract to be checked based on the knowledge graph in the at least one dimension of: text error correction, consistency check, completeness check, and logic check,

wherein performing the text error correction on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a type of an attribute; and

generating a result of the text error correction by inputting context related to the attribute in the text of the contract to be checked into a text error correction model corresponding to a preset type in response to the type of the attribute being consistent with the preset type,

wherein performing the completeness check on the text of the contract to be checked based on the knowledge graph comprises:

determining a successful checking of the completeness check of the text of the contract to be checked in response to the plurality of attributes comprising each of a plurality of term keywords in a preset completeness checking database.

2. The method according to claim 1 , wherein performing the consistency check on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a plurality of attribute values corresponding to an attribute; and

generating a result of the consistency check based on consistency of the plurality of attribute values.

3. The method according to claim 1 , wherein performing the logic check on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a logic check rule, wherein the logic check rule comprises a plurality of logic check attributes and a logical relationship between the plurality of logic check attributes;

obtaining a plurality of attribute values corresponding to the plurality of logic check attributes; and

generating a result of the logic check based on the plurality of attribute values and the logical relationship.

4. An electronic device, comprising:

at least one processor; and

a memory communicatively connected to the at least one processor; wherein,

the memory is stored with instructions executable by the at least one processor, when the instructions are performed by the at least one processor, the at least one processor is caused to perform a text checking method based on a knowledge graph, the method comprising:

obtaining a text of a contract to be checked;

establishing a knowledge graph according to the text of the contract to be checked, wherein the knowledge graph comprises a plurality of attributes and a plurality of attribute values corresponding to the plurality of attributes; and

checking the text of the contract to be checked based on the knowledge graph,

generating a target check result based on a result in at least one dimension; and

determining whether the text of the contract needs to be modified based on the target check result, comprising:

in the case that a result of one dimension is failed, determining that the target check result is failed and the contract text is modified; and in the case that all the results in the at least one dimension are successful, determining that the target check result is successful and the contract text does not need to be modified,

wherein checking the text of the contract to be checked based on the knowledge graph comprises: checking the text of the contract to be checked based on the knowledge graph in the at least one dimension of: text error correction, consistency check, completeness check, and logic check,

wherein performing the text error correction on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a type of an attribute; and

generating a result of the text error correction by inputting context related to the attribute in the text of the contract to be checked into a text error correction model corresponding to a preset type in response to the type of the attribute being consistent with the preset type,

wherein performing the completeness check on the text of the contract to be checked based on the knowledge graph comprises:

determining a successful checking of the completeness check of the text of the contract to be checked in response to the plurality of attributes comprising each of a plurality of term keywords in a preset completeness checking database.

5. The electronic device according to claim 4 , wherein performing the consistency check on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a plurality of attribute values corresponding to an attribute; and

generating a result of the consistency check based on consistency of the plurality of attribute values.

6. The electronic device according to claim 4 , wherein performing the logic check on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a logic check rule, wherein the logic check rule comprises a plurality of logic check attributes and a logical relationship between the plurality of logic check attributes;

obtaining a plurality of attribute values corresponding to the plurality of logic check attributes; and

generating a result of the logic check based on the plurality of attribute values and the logical relationship.

7. A non-transitory computer readable storage medium stored with computer instructions, wherein, the computer instructions are configured to cause a computer to perform a text checking method based on a knowledge graph, the method comprising:

obtaining a text of a contract to be checked;

establishing a knowledge graph according to the text of the contract to be checked, wherein the knowledge graph comprises a plurality of attributes and a plurality of attribute values corresponding to the plurality of attributes; and

checking the text of the contract to be checked based on the knowledge graph,

generating a target check result based on a result in at least one dimension; and

determining whether the text of the contract needs to be modified based on the target check result, comprising:

in the case that a result of one dimension is failed, determining that the target check result is failed and the contract text is modified; and in the case that all the results in the at least one dimension are successful, determining that the target check result is successful and the contract text does not need to be modified,

wherein checking the text of the contract to be checked based on the knowledge graph comprises: checking the text of the contract to be checked based on the knowledge graph in the at least one dimension of: text error correction, consistency check, completeness check, and logic check,

wherein performing the text error correction on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a type of an attribute; and

generating a result of the text error correction by inputting context related to the attribute in the text of the contract to be checked into a text error correction model corresponding to a preset type in response to the type of the attribute being consistent with the preset type,

wherein performing the completeness check on the text of the contract to be checked based on the knowledge graph comprises:

determining a successful checking of the completeness check of the text of the contract to be checked in response to the plurality of attributes comprising each of a plurality of term keywords in a preset completeness checking database.

8. The non-transitory computer readable storage medium according to claim 7 , wherein performing the consistency check on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a plurality of attribute values corresponding to an attribute; and

generating a result of the consistency check based on consistency of the plurality of attribute values.

9. The non-transitory computer readable storage medium according to claim 7 , wherein performing the logic check on the text of the contract to be checked based on the knowledge graph comprises:

obtaining a logic check rule, wherein the logic check rule comprises a plurality of logic check attributes and a logical relationship between the plurality of logic check attributes;

obtaining a plurality of attribute values corresponding to the plurality of logic check attributes; and

generating a result of the logic check based on the plurality of attribute values and the logical relationship.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jul 2, 2025
From: YANG, CHEN; SUN, ZHUO
To: BEIJING BAIDU NETCOM SCIENCE TECHNOLOGY CO., LTD.
Reel/Frame 071594/0826 →
Priority Claims (1)
CN 202111005347.1 · Aug 30, 2021 · national
Continuity (1)
Related Publication 20220391598A1 · Dec 8, 2022
References Cited (21)
US 11087219B1 · Brecque · 2021 [cited by applicant]
US 20170124497A1 · Crabtree et al. · 2017 [cited by applicant]
US 20180052884A1 · Kale et al. · 2018 [cited by applicant]
US 20190171944A1 · Lu · 2019 [cited by examiner]
US 20210158463A1 · Madl · 2021 [cited by examiner]
US 20210383247A1 · Brecque · 2021 [cited by examiner]
US 20220350690A1 · Sun · 2022 [cited by examiner]
CN 109190098A · 2019 [cited by applicant]
CN 109344382A · 2019 [cited by applicant]
CN 110442872A · 2019 [cited by applicant]
CN 112396437A · 2021 [cited by applicant]
CN 112883197A · 2021 [cited by applicant]
CN 113312902A · 2021 [cited by applicant]
CN 115081431A · 2022 [cited by examiner]
English translation of the above CN-115081431-A, Sep. 20, 2020. [cited by examiner]
Liu, Zhibin, et al. “Knowledge aware conversation generation with explainable reasoning over augmented graphs.” arXiv preprint arXiv:1903.10245 (2019). (Year: 2019). [cited by examiner]
Office Action issued for corresponding Chinese patent application No. 202111005347.1, mailed May 23, 2022 (15 pages). [cited by applicant]
Second Office Action issued for corresponding Chinese patent application No. 202111005347.1, mailed Jul. 29, 2022 (30 pages). [cited by applicant]
Bao, et al; “Customs declaration and inspection certificate operation manual Customs document;” Edition Library of China CP Data Nuclear Science, No. 098993; 2013 (10 pages). [cited by applicant]
Extended European Search Report issued in European Application No. 22191625.7, mailed on Jan. 19, 2023 (7 pages). [cited by applicant]
Mario Mezzanzanica et al.; “A Model-Based Approach for Developing Data Cleansing Solutions;” ACM Journal of Data and Information Quality; vol. 5; No. 4; Article 13; Feb. 2015; pp. 1-28 (28 pages). [cited by applicant]