A Method And System For Refinement Of Indic Language Translation

Tata Consultancy Services Limited
Granted App No 201621011932 Patent 475530 Computer Science

Patent information

IP India · live record
Application number
201621011932
Filing date
04 April 2016
Status
Granted
Legal status
Inforce
Publication number
40/2017
Publication type
INA
Invention field
Computer Science
Patent number
475530
Grant date
30 November 2023
Next renewal due
04 April 2027

Abstract

A method and system is provided for suggesting context sensitive grammatical error correction by generating one or more word forms at real time from a source text to a translated text used for the refinement of Indic language translation. A translated text and a source text is provided as an input which is then parsed into word units based on the space between each word. The word units are further processed to generate a set of most relevant word forms at run time based upon the application of rules like part of speech, root form ending, context recognition and other factors. The disclosure can accept manual and machine translated text as input and can be used to refine the language translations through a browser based user interface, including mobiles. The disclosure also explains method for context sensitive corrections suggestions provided to the user which are automatically determined in real-time.

Applicants

1
India
TATA CONSULTANCY SERVICES LIMITED Nirmal Building, 9th Floor, Nariman Point, Mumbai - 400021, Maharashtra, India

Inventors

6
India
JOSHI, ShrikantTata Consultancy Services Limited Tata Research Development & Design Centre, 54-B, Hadapsar Industrial Estate, Hadapsar, Pune - 411 013 Maharashtra, India
India
KALRA, KanikaTata Consultancy Services Limited Tata Research Development & Design Centre, 54-B, Hadapsar Industrial Estate, Hadapsar, Pune - 411 013 Maharashtra, India
India
SRIRAMAN, AnandTata Consultancy Services Limited Tata Research Development & Design Centre, 54-B, Hadapsar Industrial Estate, Hadapsar, Pune - 411 013 Maharashtra, India
India
KARANDE, Shirish SubhashTata Consultancy Services Limited Tata Research Development & Design Centre, 54-B, Hadapsar Industrial Estate, Hadapsar, Pune - 411 013 Maharashtra, India
India
LODHA, Sachin PremsukhTata Consultancy Services Limited Tata Research Development & Design Centre, 54-B, Hadapsar Industrial Estate, Hadapsar, Pune - 411 013 Maharashtra, India
India
KUMAR, RahulTata Consultancy Services Limited Tata Research Development & Design Centre, 54-B, Hadapsar Industrial Estate, Hadapsar, Pune - 411 013 Maharashtra, India

Claims

10
1 A computer implemented method for suggesting context sensitive grammatical error corrections to an output text of a translation, the method comprising processor implemented steps of: receiving, a source text and a translated text as an input by a data collector module; parsing the source text and the translated text into word units using a tokenizer module; aligning each of the word units of translated text with the corresponding word unit of the source text using an aligner module; tagging each of the word units of translated text and source text according to at least one factor wherein the at least one factor comprises a grammatical category, a gender, a number and a case using a morphological analyzer module; performing a linguistic processing on the tagged word units using a language engine component, wherein the linguistic processing comprises: extracting morphosyntactic rules of a target language by using the synchronous context free grammar technique of a rule extractor module, and generating one or more word forms of the word units of translated text considering the tag of the word units using a morphological generator module; and suggesting, using an user interface, the generated one or more word forms, wherein the generated one or more word forms to be displayed in the form of drop-down lists.
2 The method as claimed in claim 1, wherein the aligning further comprises comparing the word units of the translated text with the source text to determine which word in the translated text corresponds to which word in the source text.
3 The method as claimed in claim 1, wherein the extracting morphosyntactic rule further comprises extracting and associating at least one morphological rule, to the tagged word unit.
4 The method as claimed in claim 1, wherein the morphosyntactic rule comprises of associating the tagged word units to at least one of a extracted feature, wherein the extracted feature further comprises at least one of a root form ending, part of speech, numerals, case and gender.
5 The method as claimed in claim 1 ,wherein the generating one or more word forms further comprises providing alternate syntactical word forms, wherein the word forms correspond to change in the extracted feature of the word units at run time.
6 The method as claimed in claim 5, wherein the generating one or more word forms further comprises: extracting an interrelationship of the said word units to generate a syntactical set of word forms; tagging the said word units to a second factor based on the interrelationship; comparing the change between the second factor and previous factor of the word units; and identifying and associating the most appropriate rule to the tagged word units.
7 A system for suggesting context sensitive grammatical error corrections to an output text of a translation, the system comprising: a processor; a memory communicatively coupled with the processor, the memory comprising: a pre-processor component to receive a source text and a translated text as an input, wherein the pre-processor component comprises: a data collector module to receive the source text and a translated text; a tokenizer module to parse the source text and translated text into word units; an aligner module to align the tokenized word units of translated text corresponding to word units of source text and sending an alignment information to a morphological generator module; a morphological analyzer module generates output by analyzing and tagging the word units of source text and translated text according to at least one of a grammatical category, a gender, a number and a case; a language engine component to perform a linguistic processing of the received output of the pre-processor component, wherein the language engine component comprises: a rule engine to create and store the morphosyntactic rules; a rule extractor module to extract morphosyntactic rules of a target language by using the synchronous context free grammar technique; a morphological generator module to generate one or more word forms of the word units of translated text considering the tag of the word units and alignment information sent from the aligner module; and an user interface to suggest the generated one or more word forms, wherein the word forms to be displayed in the form of drop-down lists.
8 The system claimed in claim 7, wherein the output text of translation comprises a translated text output from manual or automated translation.
9 The system as claimed in claim 7, wherein the rules created by the rule engine module are based on part of speech and root form ending of the word units.
10 The system as claimed in claim 7, wherein the rule engine further configured to perform: a comparison of a grammatically correct word forms and incorrect word forms of the target language to determine application of the morphological rule; storing of the determined morphological rules; and suggesting the list of one or more word forms subsequent to the comparison through the user interface as a drop-down list. , Description:As Attahced

Specification

Show full textCollapse

Claims:1. A computer implemented method for suggesting context sensitive grammatical error corrections to an output text of a translation, the method comprising processor implemented steps of: receiving, a source text and a translated text as an input by a data collector module; parsing the source text and the translated text into word units using a tokenizer module; aligning each of the word units of translated text with the corresponding word unit of the source text using an aligner module; tagging each of the word units of translated text and source text according to at least one factor wherein the at least one factor comprises a grammatical category, a gender, a number and a case using a morphological analyzer module; performing a linguistic processing on the tagged word units using a language engine component, wherein the linguistic processing comprises: extracting morphosyntactic rules of a target language by using the synchronous context free grammar technique of a rule extractor module, and generating one or more word forms of the word units of translated text considering the tag of the word units using a morphological generator module; and suggesting, using an user interface, the generated one or more word forms, wherein the generated one or more word forms to be displayed in the form of drop-down lists.

2. The method as claimed in claim 1, wherein the aligning further comprises comparing the word units of the translated text with the source text to determine which word in the translated text corresponds to which word in the source text.

3. The method as claimed in claim 1, wherein the extracting morphosyntactic rule further comprises extracting and associating at least one morphological rule, to the tagged word unit.

4. The method as claimed in claim 1, wherein the morphosyntactic rule comprises of associating the tagged word units to at least one of a extracted feature, wherein the extracted feature further comprises at least one of a root form ending, part of speech, numerals, case and gender.

5. The method as claimed in claim 1 ,wherein the generating one or more word forms further comprises providing alternate syntactical word forms, wherein the word forms correspond to change in the extracted feature of the word units at run time.

6. The method as claimed in claim 5, wherein the generating one or more word forms further comprises: extracting an interrelationship of the said word units to generate a syntactical set of word forms; tagging the said word units to a second factor based on the interrelationship; comparing the change between the second factor and previous factor of the word units; and identifying and associating the most appropriate rule to the tagged word units. 7. A system for suggesting context sensitive grammatical error corrections to an output text of a translation, the system comprising: a processor; a memory communicatively coupled with the processor, the memory comprising: a pre-processor component to receive a source text and a translated text as an input, wherein the pre-processor component comprises: a data collector module to receive the source text and a translated text; a tokenizer module to parse the source text and translated text into word units; an aligner module to align the tokenized word units of translated text corresponding to word units of source text and sending an alignment information to a morphological generator module; a morphological analyzer module generates output by analyzing and tagging the word units of source text and translated text according to at least one of a grammatical category, a gender, a number and a case; a language engine component to perform a linguistic processing of the received output of the pre-processor component, wherein the language engine component comprises: a rule engine to create and store the morphosyntactic rules; a rule extractor module to extract morphosyntactic rules of a target language by using the synchronous context free grammar technique; a morphological generator module to generate one or more word forms of the word units of translated text considering the tag of the word units and alignment information sent from the aligner module; and an user interface to suggest the generated one or more word forms, wherein the word forms to be displayed in the form of drop-down lists.

8. The system claimed in claim 7, wherein the output text of translation comprises a translated text output from manual or automated translation.

9. The system as claimed in claim 7, wherein the rules created by the rule engine module are based on part of speech and root form ending of the word units.

10. The system as claimed in claim 7, wherein the rule engine further configured to perform: a comparison of a grammatically correct word forms and incorrect word forms of the target language to determine application of the morphological rule; storing of the determined morphological rules; and

suggesting the list of one or more word forms subsequent to the comparison through the user interface as a drop-down list. , Description:As Attahced

Documents

Sign in to open
Orders
Section
Application documents
1 Form 5 [04-04-2016(online)].pdf 04 Apr 2016
2 Form 3 [04-04-2016(online)].pdf 04 Apr 2016
3 Form 18 [04-04-2016(online)].pdf 04 Apr 2016
4 Drawing [04-04-2016(online)].pdf 04 Apr 2016
5 Description(Complete) [04-04-2016(online)].pdf 04 Apr 2016
6 Other Patent Document [16-09-2016(online)].pdf 16 Sep 2016
7 ABSTRACT.jpg 11 Aug 2018
8 201621011932-POWER OF ATTORNEY-(02-5-2016).pdf 11 Aug 2018
9 201621011932-Form 1-220916.pdf 11 Aug 2018
10 201621011932-Correspondence-220916.pdf 11 Aug 2018
11 201621011932-CORRESPONDENCE-(02-5-2016).pdf 11 Aug 2018
12 201621011932-FER.pdf 30 Apr 2020
13 201621011932-OTHERS [21-10-2020(online)].pdf 21 Oct 2020
14 201621011932-FER_SER_REPLY [21-10-2020(online)].pdf 21 Oct 2020
15 201621011932-DRAWING [21-10-2020(online)].pdf 21 Oct 2020
16 201621011932-COMPLETE SPECIFICATION [21-10-2020(online)].pdf 21 Oct 2020
17 201621011932-CLAIMS [21-10-2020(online)].pdf 21 Oct 2020
18 201621011932-US(14)-HearingNotice-(HearingDate-02-11-2023).pdf 11 Oct 2023
19 201621011932-Correspondence to notify the Controller [17-10-2023(online)].pdf 17 Oct 2023
20 201621011932-FORM-26 [31-10-2023(online)].pdf 31 Oct 2023
21 201621011932-Written submissions and relevant documents [17-11-2023(online)].pdf 17 Nov 2023
22 201621011932-PatentCertificate30-11-2023.pdf 30 Nov 2023
23 201621011932-IntimationOfGrant30-11-2023.pdf 30 Nov 2023
Search strategy
1US20150046147A1E_28-04-2020.pdf
2US20120316862A1E_28-04-2020.pdf
32020-04-2816-14-17E_28-04-2020.pdf

ERegister / renewals

Inforce
3rd Renewed 11 Jan 2024 Period 04 Apr 2018 – 04 Apr 2019 · CBR 1685 · Certificate 3573 ₹4,000
4th Renewed 11 Jan 2024 Period 04 Apr 2019 – 04 Apr 2020 · CBR 1685 · Certificate 3574 ₹4,000
5th Renewed 11 Jan 2024 Period 04 Apr 2020 – 04 Apr 2021 · CBR 1685 · Certificate 3575 ₹4,000
6th Renewed 11 Jan 2024 Period 04 Apr 2021 – 04 Apr 2022 · CBR 1685 · Certificate 3576 ₹4,000
7th Renewed 11 Jan 2024 Period 04 Apr 2022 – 04 Apr 2023 · CBR 1685 · Certificate 3577 ₹12,000
8th Renewed 11 Jan 2024 Period 04 Apr 2023 – 04 Apr 2024 · CBR 1685 · Certificate 3578 ₹12,000
9th Renewed 11 Jan 2024 Period 04 Apr 2024 – 04 Apr 2025 · CBR 1685 · Certificate 3579 ₹12,000
10th Renewed 02 Apr 2025 Period 04 Apr 2025 – 04 Apr 2026 · CBR 13111 · Certificate 15159 ₹12,000
11th Renewed 01 Apr 2026 Period 04 Apr 2026 – 04 Apr 2027 · CBR 14263 · Certificate 14980 ₹24,000