System And Method For Annotation Of Tokens For Natural Language Processing

Wipro Limited
Granted App No 201941009531 Patent 543443 Computer Science

Patent information

IP India · live record
Application number
201941009531
Filing date
12 March 2019
Status
Granted
Legal status
Inforce
Publication number
38/2020
Publication type
INA
Invention field
Computer Science
Patent number
543443
Grant date
27 June 2024
Next renewal due
12 March 2026Overdue

Abstract

This disclosure relates to method and system for annotating tokens for natural language processing (NLP). In one embodiment, the method may include segmenting a plurality of corpus based on each of a plurality of instances, deriving a plurality of entities for each of the plurality of instances based on at least one of a machine learning technique and a deep learning technique, determining a word vector for each of the plurality of entities associated with each of the plurality of instances, and labelling a plurality of tokens for each of the plurality of instances. It should be noted that the plurality of tokens associated with the plurality of entities may be identified based on a frequency of each of the plurality of entities. FIG. 1

Applicants

1
India
WIPRO LIMITED Doddakannelli, Sarjapur Road, Bangalore 560035, Karnataka, India.

Inventors

2
India
RISHAV DAS33/1, Nandi Bagan Bye Lane, P.O. Salkia, P.S. Golabari, Howrah 711106, West Bengal, India.
India
SOURAV MUDIPahalanpur, Burdwan, Madhabdihi 713427, West Bengal, India.

Claims

12
1 A method of annotating tokens for natural language processing, the method comprising: segmenting, by an automatic annotation device, a plurality of corpus based on each of a plurality of instances; deriving, by the automatic annotation device, a plurality of entities for each of the plurality of instances based on at least one of a machine learning technique and a deep learning technique; determining, by the automatic annotation device, a word vector for each of the plurality of entities associated with each of the plurality of instances; and labelling, by the automatic annotation device, a plurality of tokens for each of the plurality of instances, wherein the plurality of tokens associated with the plurality of entities are identified based on a frequency of each of the plurality of entities.
2 The method of claim 1, further comprising: receiving a plurality of dataset for the plurality of instances from at least one of a website, a portable document format (PDF), a research paper, a journal, or an article; and determining the plurality of corpus by clubbing each of the plurality of dataset.
3 The method of claim 1, further comprising: determining a pattern for each of the plurality of tokens based on an alphabetic embedding; and identifying a matching tokens of the plurality of tokens based on the pattern, wherein the matching tokens is used to identify at least one of a rhythm, a poetry, or a prose.
4 The method of claim 1, wherein labelling the plurality of tokens further comprises: measuring an accuracy of each of the plurality of tokens for each of the plurality of instances using a confusion matrix; and re-labelling the plurality of tokens for each of the plurality of instances when the accuracy is below an accuracy threshold, wherein the accuracy threshold can be scaled up or scaled down based on a mistake performed by a neural network.
5 The method of claim 4, further comprising training the neural network based on the plurality of tokens when the accuracy of each of the plurality of token is above the accuracy threshold.
6 The method of claim 5, further comprising: determining an accuracy value of each of the plurality of entities based on n-fold cross validation technique; and re-labelling each of the plurality of tokens when the accuracy value is not in a predefined range.
7 A system for annotating tokens for natural language processing, the system comprising: an automatic annotation device comprising at least one processor and a computer-readable medium storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising: segmenting a plurality of corpus based on each of a plurality of instances; deriving a plurality of entities for each of the plurality of instances based on at least one of a machine learning technique and a deep learning technique; determining a word vector for each of the plurality of entities associated with each of the plurality of instances; and labelling a plurality of tokens for each of the plurality of instances, wherein the plurality of tokens associated with the plurality of entities are identified based on a frequency of each of the plurality of entities.
8 The system of claim 7, wherein the operations further comprise: receiving a plurality of dataset for the plurality of instances from at least one of a website, a portable document format (PDF), a research paper, a journal, or an article; and determining the plurality of corpus by clubbing each of the plurality of dataset.
9 The system of claim 7, wherein the operations further comprise: determining a pattern for each of the plurality of tokens based on an alphabetic embedding; and identifying a matching tokens of the plurality of tokens based on the pattern, wherein the matching tokens is used to identify at least one of a rhythm, a poetry, or a prose.
10 The system of claim 7, wherein labelling the plurality of tokens further comprises: measuring an accuracy of each of the plurality of tokens for each of the plurality of instances using a confusion matrix; and re-labelling the plurality of tokens for each of the plurality of instances when the accuracy is below an accuracy threshold, wherein the accuracy threshold can be scaled up or scaled down based on a mistake performed by a neural network.
11 The system of claim 10, wherein the operations further comprise training the neural network based on the plurality of tokens when the accuracy of each of the plurality of token is above the accuracy threshold.
12 The system of claim 11, wherein the operations further comprise: determining an accuracy value of each of the plurality of entities based on n-fold cross validation technique; and re-labelling each of the plurality of tokens when the accuracy value is not in a predefined range. Dated 12th day of March, 2019 R Ramya Rao Of K&S Partners Agent for the Applicant IN/PA-1607 , Description:TECHNICAL FIELD [001] This disclosure relates generally to natural language processing (NLP), and more particularly to method and system for annotating tokens for NLP applications.

Specification

Show full textCollapse

Claims:WE CLAIM: 1. A method of annotating tokens for natural language processing, the method comprising: segmenting, by an automatic annotation device, a plurality of corpus based on each of a plurality of instances; deriving, by the automatic annotation device, a plurality of entities for each of the plurality of instances based on at least one of a machine learning technique and a deep learning technique; determining, by the automatic annotation device, a word vector for each of the plurality of entities associated with each of the plurality of instances; and labelling, by the automatic annotation device, a plurality of tokens for each of the plurality of instances, wherein the plurality of tokens associated with the plurality of entities are identified based on a frequency of each of the plurality of entities.

2. The method of claim 1, further comprising: receiving a plurality of dataset for the plurality of instances from at least one of a website, a portable document format (PDF), a research paper, a journal, or an article; and determining the plurality of corpus by clubbing each of the plurality of dataset.

3. The method of claim 1, further comprising: determining a pattern for each of the plurality of tokens based on an alphabetic embedding; and identifying a matching tokens of the plurality of tokens based on the pattern, wherein the matching tokens is used to identify at least one of a rhythm, a poetry, or a prose.

4. The method of claim 1, wherein labelling the plurality of tokens further comprises: measuring an accuracy of each of the plurality of tokens for each of the plurality of instances using a confusion matrix; and re-labelling the plurality of tokens for each of the plurality of instances when the accuracy is below an accuracy threshold, wherein the accuracy threshold can be scaled up or scaled down based on a mistake performed by a neural network.

5. The method of claim 4, further comprising training the neural network based on the plurality of tokens when the accuracy of each of the plurality of token is above the accuracy threshold.

6. The method of claim 5, further comprising: determining an accuracy value of each of the plurality of entities based on n-fold cross validation technique; and re-labelling each of the plurality of tokens when the accuracy value is not in a predefined range.

7. A system for annotating tokens for natural language processing, the system comprising: an automatic annotation device comprising at least one processor and a computer-readable medium storing instructions that, when executed by the at least one processor, cause the at least one processor to perform operations comprising: segmenting a plurality of corpus based on each of a plurality of instances; deriving a plurality of entities for each of the plurality of instances based on at least one of a machine learning technique and a deep learning technique; determining a word vector for each of the plurality of entities associated with each of the plurality of instances; and labelling a plurality of tokens for each of the plurality of instances, wherein the plurality of tokens associated with the plurality of entities are identified based on a frequency of each of the plurality of entities.

8. The system of claim 7, wherein the operations further comprise: receiving a plurality of dataset for the plurality of instances from at least one of a website, a portable document format (PDF), a research paper, a journal, or an article; and determining the plurality of corpus by clubbing each of the plurality of dataset.

9. The system of claim 7, wherein the operations further comprise: determining a pattern for each of the plurality of tokens based on an alphabetic embedding; and identifying a matching tokens of the plurality of tokens based on the pattern, wherein the matching tokens is used to identify at least one of a rhythm, a poetry, or a prose.

10. The system of claim 7, wherein labelling the plurality of tokens further comprises: measuring an accuracy of each of the plurality of tokens for each of the plurality of instances using a confusion matrix; and re-labelling the plurality of tokens for each of the plurality of instances when the accuracy is below an accuracy threshold, wherein the accuracy threshold can be scaled up or scaled down based on a mistake performed by a neural network.

11. The system of claim 10, wherein the operations further comprise training the neural network based on the plurality of tokens when the accuracy of each of the plurality of token is above the accuracy threshold.

12. The system of claim 11, wherein the operations further comprise: determining an accuracy value of each of the plurality of entities based on n-fold cross validation technique; and re-labelling each of the plurality of tokens when the accuracy value is not in a predefined range.

Dated 12th day of March, 2019

R Ramya Rao Of K&S Partners Agent for the Applicant IN/PA-1607 , Description:TECHNICAL FIELD

[001] This disclosure relates generally to natural language processing (NLP), and more particularly to method and system for annotating tokens for NLP applications.

Documents

Sign in to open
Application documents
1 201941009531-STATEMENT OF UNDERTAKING (FORM 3) [12-03-2019(online)].pdf 12 Mar 2019
2 201941009531-REQUEST FOR EXAMINATION (FORM-18) [12-03-2019(online)].pdf 12 Mar 2019
3 201941009531-POWER OF AUTHORITY [12-03-2019(online)].pdf 12 Mar 2019
4 201941009531-FORM 18 [12-03-2019(online)].pdf 12 Mar 2019
5 201941009531-FORM 1 [12-03-2019(online)].pdf 12 Mar 2019
6 201941009531-DRAWINGS [12-03-2019(online)].pdf 12 Mar 2019
7 201941009531-DECLARATION OF INVENTORSHIP (FORM 5) [12-03-2019(online)].pdf 12 Mar 2019
8 201941009531-COMPLETE SPECIFICATION [12-03-2019(online)].pdf 12 Mar 2019
9 201941009531-Request Letter-Correspondence [13-03-2019(online)].pdf 13 Mar 2019
10 201941009531-Power of Attorney [13-03-2019(online)].pdf 13 Mar 2019
11 201941009531-Form 1 (Submitted on date of filing) [13-03-2019(online)].pdf 13 Mar 2019
12 abstract 201941009531.jpg 14 Mar 2019
13 201941009531-Proof of Right (MANDATORY) [10-09-2019(online)].pdf 10 Sep 2019
14 Correspondence by Agent_Form 1_16-09-2019.pdf 16 Sep 2019
15 201941009531-PETITION UNDER RULE 137 [30-09-2021(online)].pdf 30 Sep 2021
16 201941009531-OTHERS [30-09-2021(online)].pdf 30 Sep 2021
17 201941009531-OTHERS [30-09-2021(online)]-1.pdf 30 Sep 2021
18 201941009531-FORM 3 [30-09-2021(online)].pdf 30 Sep 2021
19 201941009531-FER_SER_REPLY [30-09-2021(online)].pdf 30 Sep 2021
20 201941009531-FER_SER_REPLY [30-09-2021(online)]-1.pdf 30 Sep 2021
21 201941009531-DRAWING [30-09-2021(online)].pdf 30 Sep 2021
22 201941009531-DRAWING [30-09-2021(online)]-1.pdf 30 Sep 2021
23 201941009531-COMPLETE SPECIFICATION [30-09-2021(online)].pdf 30 Sep 2021
24 201941009531-COMPLETE SPECIFICATION [30-09-2021(online)]-1.pdf 30 Sep 2021
25 201941009531-CLAIMS [30-09-2021(online)].pdf 30 Sep 2021
26 201941009531-CLAIMS [30-09-2021(online)]-1.pdf 30 Sep 2021
27 201941009531-FER.pdf 17 Oct 2021
28 201941009531-US(14)-HearingNotice-(HearingDate-02-11-2023).pdf 05 Oct 2023
29 201941009531-POA [10-10-2023(online)].pdf 10 Oct 2023
30 201941009531-FORM 13 [10-10-2023(online)].pdf 10 Oct 2023
31 201941009531-Correspondence to notify the Controller [10-10-2023(online)].pdf 10 Oct 2023
32 201941009531-AMENDED DOCUMENTS [10-10-2023(online)].pdf 10 Oct 2023
33 201941009531-FORM-26 [02-11-2023(online)].pdf 02 Nov 2023
34 201941009531-Written submissions and relevant documents [17-11-2023(online)].pdf 17 Nov 2023
35 201941009531-FORM 3 [17-11-2023(online)].pdf 17 Nov 2023
36 201941009531-PatentCertificate27-06-2024.pdf 27 Jun 2024
37 201941009531-IntimationOfGrant27-06-2024.pdf 27 Jun 2024
Search strategy
12021-04-0511-42-57E_05-04-2021.pdf

ERegister / renewals

Inforce
3rd Renewed 26 Sep 2024 Period 12 Mar 2021 – 12 Mar 2022 · CBR 58979 · Certificate 151122 ₹4,000
4th Renewed 26 Sep 2024 Period 12 Mar 2022 – 12 Mar 2023 · CBR 58979 · Certificate 151123 ₹4,000
5th Renewed 26 Sep 2024 Period 12 Mar 2023 – 12 Mar 2024 · CBR 58979 · Certificate 151124 ₹4,000
6th Renewed 26 Sep 2024 Period 12 Mar 2024 – 12 Mar 2025 · CBR 58979 · Certificate 151125 ₹4,000
7th Renewed 07 Mar 2025 Period 12 Mar 2025 – 12 Mar 2026 · CBR 13387 · Certificate 17674 ₹12,000