DETAILED ACTION
Introduction
This office action is in response to Applicant’s submission filed on July 17, 2024.
Claims 1-20 are pending in the application. As such, claims 1-20 have been examined.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Drawings
The drawings were received on July 17, 2024. These drawings have been accepted and considered by the Examiner.
Claim Objections
Claim 19 is objected to because of the following informalities:
Claim 19 line 3 reads “one or more words of the original passage of text of the given graded sentence pair”. Examiner believes this to be a clerical error and it is intended to read “substituting one or more words of the original passage of text of the given graded sentence pair” in order to be consistent with the rest of the claims
Appropriate correction is required.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1-20 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-20 of U.S. Patent No. 12,073,189. Although the claims at issue are not identical, they are not patentably distinct from each other because they both claim they both claim a system generating synthetic sentence pairs, using a plurality of training signals to indicate how the given synthetic sentence pair was generated and further having humans grade the sentence pairs and a student network to predict a grade.
Specifically, the following correspondence for nonstatutory double patenting applies between the claims of U.S. Patent No. 12,073,189 and the claims of the present application, where the claim of the present application is given first, and the corresponding claim of U.S. Patent No. 12,073,189 is given second in the following claim pairs: 1, 1; 2, 1; 3, 7; 4, 8; 5, 5; 6, 6; 7, 2; 8, 3; 9, 4; 10, 9; 11, 10; 12, 13; 13, 13; 14, 18; 15, 8; 16, 17; 17, 14; 18, 2; 19, 14; and 20, 15.
Claims 1-20 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-20 of U.S. Patent No. 12,073,189. Note: Mapping below is demonstrating a nonstatutory double patenting of the instant application against issued patent of U.S. Patent No. 12,073,189, however, issued patent of US11551002, US11704506, US11875115 and US12223273 have similar claims as the instant application, and as such a timely filed terminal disclaimer for these issued patents will be required to overcome the nonstatutory double patenting rejection.
The detailed correspondence between the claims of the present Application, and those of U.S. Patent No. 12,073,189 is as follows:
Instant claims 18/775,058
Patent 12,073,189
1. A method, comprising: obtaining, by one or more processors of a processing system, for each given synthetic sentence pair of a plurality of synthetic sentence pairs: a first training signal based on how the given synthetic sentence pair was generated; and a second training signal based on a model prediction regarding a likelihood of how the given synthetic sentence pair could have been generated; generating, by the one or more processors, a plurality of graded sentence pairs, each graded sentence pair of the plurality of graded sentence pairs comprising an original passage of text, a modified passage of text and a grade generated based on the original passage of text and the modified passage of text; and training, by the one or more processors, a student network to predict, for each graded sentence pair in the plurality of graded sentence pairs, the grade generated by a neural network.
1. A method of training a neural network, comprising: generating, by one or more processors of a processing system, for each given synthetic sentence pair of a plurality of synthetic sentence pairs comprising a first passage of text and a second passage of text: a first training signal of a plurality of training signals based on whether the given synthetic sentence pair was generated using backtranslation; and one or more second training signals of the plurality of training signals based on a model prediction regarding a likelihood that one of the first passage of text or the second passage of text of the given synthetic sentence pair could have been generated by backtranslating the other one of the first passage of text or the second passage of text of the given synthetic sentence pair; generating, by the one or more processors, a plurality of graded sentence pairs, each graded sentence pair of the plurality of graded sentence pairs comprising an original passage of text and a modified passage of text and a grade generated by a trained neural network based on the original passage of text and the modified passage of text; and training, by the one or more processors, a student network to predict, for each given graded sentence pair in the plurality of graded sentence pairs, the grade generated by the neural network.
2. The method of claim 1, wherein the given synthetic sentence pair comprises a first passage of text and a second passage of text.
1. A method of training a neural network, comprising: generating, by one or more processors of a processing system, for each given synthetic sentence pair of a plurality of synthetic sentence pairs comprising a first passage of text and a second passage of text:
3. The method of claim 2, further comprising obtaining, for each given synthetic sentence pair of the plurality of synthetic sentence pairs, one or more third training signals based on one or more scores.
7. The method of claim 1, further comprising generating, for each given synthetic sentence pair of the plurality of synthetic sentence pairs, one or more third training signals of the plurality of training signals based on one or more scores.
4. The method of claim 3, wherein the one or more scores are generated by comparing the first passage of text to the second passage of text using one or more metrics.
8. The method of claim 7, wherein the one or more scores are generated by comparing the first passage of text to the second passage of text using one or more automatic metrics.
5. The method of claim 2, further comprising generating the plurality of synthetic sentence pairs.
5. The method of claim 1, further comprising generating the plurality of synthetic sentence pairs.
6. The method of claim 5, wherein generating the plurality of synthetic sentence pairs comprises substituting one or more words of the first passage of text to create the second passage of text.
6. The method of claim 5, wherein generating the plurality of synthetic sentence pairs comprises substituting one or more words of the first passage of text to create the second passage of text.
7. The method of claim 1, wherein the plurality of synthetic sentence pairs comprises text in a plurality of different languages, and the plurality of graded sentence pairs comprises text in only a subset of the plurality of different languages.
2. The method of claim 1, wherein the plurality of synthetic sentence pairs comprises text in a plurality of different languages, and the plurality of graded sentence pairs comprises text in only a subset of the plurality of different languages.
8. The method of claim 1, wherein generating the plurality of graded sentence pairs comprises, for each given graded sentence pair of a first subset of the graded sentence pairs, substituting one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair.
3. The method of claim 1, wherein generating the plurality of graded sentence pairs comprises, for each given graded sentence pair of a first subset of the graded sentence pairs, substituting one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair.
9. The method of claim 8, wherein generating the plurality of graded sentence pairs includes removing one or more words of the original passage of text to create the modified passage of text.
4. The method of claim 3, wherein generating the plurality of graded sentence pairs includes removing one or more words of the original passage of text to create the modified passage of text.
10. The method of claim 1, further comprising training the neural network based on each given synthetic sentence pair and the first and second training signals.
9. The method of claim 1, further comprising training the neural network based on each given synthetic sentence pair and the plurality of training signals.
11. The method of claim 1, further comprising training the neural network based on a plurality of human-graded sentence pairs.
10. The method of claim 1, wherein training the neural network is further based on a plurality of human-graded sentence pairs.
12. A processing system comprising: a memory; and one or more processors coupled to the memory and configured to: obtain, for each given synthetic sentence pair of a plurality of synthetic sentence pairs: a first training signal based on how the given synthetic sentence pair was generated; and a second training signal based on a model prediction regarding a likelihood of how the given synthetic sentence pair could have been generated; generate a plurality of graded sentence pairs, each graded sentence pair of the plurality of graded sentence pairs comprising an original passage of text, a modified passage of text and a grade generated based on the original passage of text and the modified passage of text; and train a student network to predict, for each graded sentence pair in the plurality of graded sentence pairs, the grade generated by a neural network.
13. A processing system comprising: a memory; and one or more processors coupled to the memory and configured to: generate, for each given synthetic sentence pair of a plurality of synthetic sentence pairs comprising a first passage of text and a second passage of text: a first training signal of a plurality of training signals based on whether the given synthetic sentence pair was generated using backtranslation; and one or more second training signals of the plurality of training signals based on a model prediction regarding a likelihood that one of the first passage of text or the second passage of text of the given synthetic sentence pair could have been generated by backtranslating the other one of the first passage of text or the second passage of text of the given synthetic sentence pair; generate a plurality of graded sentence pairs, each graded sentence pair of the plurality of graded sentence pairs comprising an original passage of text and a modified passage of text and a grade generated by a trained neural network based on the original passage of text and the modified passage of text; and train a student network to predict, for each given graded sentence pair in the plurality of graded sentence pairs, the grade generated by the trained neural network.
13. The processing system of claim 12, wherein the given synthetic sentence pair comprises a first passage of text and a second passage of text.
13. A processing system comprising: a memory; and one or more processors coupled to the memory and configured to: generate, for each given synthetic sentence pair of a plurality of synthetic sentence pairs comprising a first passage of text and a second passage of text:
14. The processing system of claim 13, wherein the one or more processors are further configured to generate, for each given synthetic sentence pair of the plurality of synthetic sentence pairs, one or more third training signals based on one or more scores.
18. The system of claim 13, wherein the one or more processors are further configured to generate, for each given synthetic sentence pair of the plurality of synthetic sentence pairs: one or more third training signals of the plurality of training signals based on one or more scores generated by comparing the first passage of text of the given synthetic sentence pair to the second passage of text of the given synthetic sentence pair.
15. The processing system of claim 14, wherein the one or more processors generate the one or more scores by comparing the first passage of text to the second passage of text using one or more metrics.
8. The method of claim 7, wherein the one or more scores are generated by comparing the first passage of text to the second passage of text using one or more automatic metrics.
16. The processing system of claim 13, wherein the one or more processors are further configured to generate the plurality of synthetic sentence pairs.
17. The system of claim 13, wherein the one or more processors are further configured to generate the plurality of synthetic sentence pairs.
17. The processing system of claim 13, wherein the one or more processors are further configured to substitute one or more words of the first passage of text to create the second passage of text.
14. The system of claim 13, wherein the one or more processors are further configured to, for each given graded sentence pair of a first subset of the graded sentence pairs, substitute one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair.
18. The processing system of claim 12, wherein the plurality of synthetic sentence pairs comprises text in a plurality of different languages, and the plurality of graded sentence pairs comprises text in only a subset of the plurality of different languages.
2. The method of claim 1, wherein the plurality of synthetic sentence pairs comprises text in a plurality of different languages, and the plurality of graded sentence pairs comprises text in only a subset of the plurality of different languages.
19. The processing system of claim 12, wherein the one or more processors are further configured to substitute, for each given graded sentence pair of a first subset of the graded sentence pairs, one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair.
14. The system of claim 13, wherein the one or more processors are further configured to, for each given graded sentence pair of a first subset of the graded sentence pairs, substitute one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair.
20. The processing system of claim 12, wherein the one or more processors are further configured to remove one or more words of the original passage of text to create the modified passage of text.
15. The system of claim 14, wherein the one or more processors are further configured to, for each given graded sentence pair of a second subset of the graded sentence pairs, remove one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-5, 7, 10-16 and 18 are rejected under 35 U.S.C. 103 as unpatentable over Le et al. (US Patent Pub. No. 2016/0117316), hereinafter Le, in view of Bojar et al. (US Patent Pub. No. 2020/0210772), hereinafter Bojar, in view of Song et al. (US Patent Pub. No. 2017/0060855), hereinafter Song.
Regarding claims 1 and 12, Le teaches a method and a processing system (Le in [0006] teaches This specification describes how a system implemented as one or more computer programs on one or more computers can be trained to perform, and perform, natural language translations using neural network translation models and rare word post-processing),
comprising:
[claim 12 only] a memory; and one or more processors coupled to the memory and configured to (Le in [0006] teaches using computers [which have processors and memory]:
obtaining, by one or more processors of a processing system, for each given synthetic sentence pair of a plurality of synthetic sentence pairs (Le in [0006] teaches using computers [which have processors and memory], and Le in [0014] teaches The translation system 100 receives sentences in a source natural language, e.g., a source language sentence [source language sentence maps to first passage of text] 110, and translates the source natural language sentences into target sentences in a target natural language, e.g., a target language sentence [target language sentence maps to second passage of text] 150 for the source language sentence 110 [together the source language sentence and the target language sentence maps to a synthetic sentence pair]):
Le does not teach, however Bojar teaches
a first training signal based on how the given synthetic sentence pair was generated (Bojar in [0026] teaches the next step, in which a first auxiliary translation system is trained on the corpus C, said trained first auxiliary translation system is then used to translate the corpus B from third language to source language resulting in a back-translated corpus D [sentences identified as from corpus D maps to first training signal], which is further filtered to keep only similar sentences to those contained in the noisy corpus G resulting in a synthetic parallel corpus D2);
and
generating, by the one or more processors, a plurality of graded sentence pairs, each graded sentence pair of the plurality of graded sentence pairs comprising an original passage of text, a modified passage of text and a grade generated based on the original passage of text and the modified passage of text (Bojar in [0044] teaches Training a translation model with the corpus FINAL is performed using well known approaches, such as tensor2tensor transformer and RNN—Recurrent neural network architectures. Automatic validation could be then performed using BLEU metric, Meteor, CHRF3 or other suitable automatic metric on both noisy and clean validation corpora. Alternatively as a “translation quality metric score” any scoring algorithm for evaluating the quality of translated text based on existing human translations could be used [translation quality metric score maps to grading a sentence pair, this can be applied to each passage of text to obtain graded passages of text (original passage of text and a modified passage of text) from the first and second passages of text]),
and
training, by the one or more processors, a student network to predict, for each graded sentence pair in the plurality of graded sentence pairs, the grade generated by a neural network (Bojar in [0044] teaches Training a translation model with the corpus FINAL is performed using well known approaches, such as tensor2tensor transformer and RNN—Recurrent neural network architectures. Automatic validation could be then performed using BLEU metric, Meteor, CHRF3 or other suitable automatic metric on both noisy and clean validation corpora. Alternatively as a “translation quality metric score” any scoring algorithm for evaluating the quality of translated text based on existing human translations could be used [translation quality metric score also maps to predicting a grade for a sentence pair, this can be applied to each passage of text to obtain predicted grades for passages of text (original passage of text and a modified passage of text) from the first and second passages of text]).
Bojar is considered to be analogous to the claimed invention because it is in the same field of neural network based training for translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le further in view of Bojar to allow for back-translating and comparing to human translations. Doing so would provide a system which is tolerant to noisy inputs and improve translation accuracy even for low resource language pairs (Bojar, para. 14).
Le, as modified above, does not teach, however Song teaches
a second training signal based on a model prediction regarding a likelihood of how the given synthetic sentence pair could have been generated (Song in [0131] teaches After extraction of the translation rules, the computing device may extract features of the translation rules. The features of the translation rules may include: a forward translation probability, reverse translation probability, positive vocabulary probability, and reverse vocabulary probability. In these instances, the forward translation probabilities of phrases [forward translation probabilities maps to one (or more) second training signals] refer to a translation probability of a translation of a phrase from a source language to a target language. The reverse translation probabilities of phrases [reverse translation probabilities maps to (one or) more second training signals] refer to a translation probability of a translation of a phrase from a target language to a source language. The positive vocabulary probability refers to a translation probability of a word from a source language to a target language. The reverse vocabulary probability refers to a translation probability of a translation of a word from a target language to a source language);
Song is considered to be analogous to the claimed invention because it is in the same field of neural networks used for bilingual prediction models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Song to allow for predicting the likelihood of back-translation and training the neural network. Doing so would resolve issues relating to semantic inconsistency of candidate translations (Song, para. 7).
Regarding claims 2 and 13, Le, as modified above, teaches the method and processing system of claims 1 and 12.
Le further teaches
wherein the given synthetic sentence pair comprises a first passage of text and a second passage of text (Le in [0014] teaches The translation system 100 receives sentences in a source natural language, e.g., a source language sentence [source language sentence maps to first passage of text] 110, and translates the source natural language sentences into target sentences in a target natural language, e.g., a target language sentence [target language sentence maps to second passage of text] 150 for the source language sentence 110 [together the source language sentence and the target language sentence maps to a synthetic sentence pair]).
Regarding claims 3 and 14, Le, as modified above, teaches the method and processing system of claims 2 and 13.
Le, as modified above, does not teach, however Bojar teaches
further comprising
[claim 14 only] wherein the one or more processors are further configured to generate,
obtaining, for each given synthetic sentence pair of the plurality of synthetic sentence pairs, one or more third training signals based on one or more scores (Bojar in [0044] teaches Training a translation model with the corpus FINAL is performed using well known approaches, such as tensor2tensor transformer and RNN—Recurrent neural network architectures. Automatic validation could be then performed using BLEU metric, Meteor, CHRF3 or other suitable automatic metric on both noisy and clean validation corpora. Alternatively as a “translation quality metric score” any scoring algorithm for evaluating the quality of translated text based on existing human translations could be used [translation quality metric score also maps to one or more third training signals]).
Bojar is considered to be analogous to the claimed invention because it is in the same field of neural network based training for translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Bojar to allow for generating translation quality metric scores as one or more third training signals. Doing so would provide a system which is tolerant to noisy inputs and improve translation accuracy even for low resource language pairs (Bojar, para. 14).
Regarding claims 4 and 15, Le, as modified above, teaches the method and processing system of claims 3 and 14.
Le, as modified above, does not teach, however Bojar teaches
[claim 15 only] wherein the one or more processors [generate the one or more scores]
wherein the one or more scores are generated by comparing the first passage of text to the second passage of text using one or more metrics (Bojar [0044] Training a translation model with the corpus FINAL is performed using well known approaches, such as tensor2tensor transformer and RNN—Recurrent neural network architectures. Automatic validation could be then performed using BLEU metric, Meteor, CHRF3 or other suitable automatic metric [BLEU metric, Meteor, CHRF3 maps to one or more automatic metrics] on both noisy and clean validation corpora. Alternatively as a “translation quality metric score” any scoring algorithm for evaluating the quality of translated text based on existing human translations could be used).
Bojar is considered to be analogous to the claimed invention because it is in the same field of neural network based training for translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Bojar to allow for using an automatic validation metric such as BLEU. Doing so would provide a system which is tolerant to noisy inputs and improve translation accuracy even for low resource language pairs (Bojar, para. 14).
Regarding claims 5 and 16, Le, as modified above, teaches the method and processing system of claims 2 and 13.
Le further teaches
further comprising
[claim 16 only] wherein the one or more processors are further configured to
generating the plurality of synthetic sentence pairs (Le in [0014] teaches The translation system 100 receives sentences in a source natural language, e.g., a source language sentence [source language sentence maps to first passage of text] 110, and translates the source natural language sentences into target sentences in a target natural language, e.g., a target language sentence [target language sentence maps to second passage of text] 150 for the source language sentence 110 [together the source language sentence and the target language sentence maps to a synthetic sentence pair]).
Regarding claims 7 and 18, Le, as modified above, teaches the method and processing system of claims 1 and 12.
Le, as modified above, does not teach, however Bojar teaches
wherein the plurality of synthetic sentence pairs comprises text in a plurality of different languages, and the plurality of graded sentence pairs comprises text in only a subset of the plurality of different languages (Bojar in [0026] teaches the next step, in which a first auxiliary translation system is trained on the corpus C, said trained first auxiliary translation system is then used to translate the corpus B from third language to source language resulting in a back-translated corpus D, which is further filtered to keep only similar sentences to those contained in the noisy corpus G resulting in a synthetic parallel corpus D2 [further filtered to keep only similar sentences maps to a subset]).
Bojar is considered to be analogous to the claimed invention because it is in the same field of neural network based training for translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Bojar to allow for grading only a subset of the sentence pairs. Doing so would provide a system which is tolerant to noisy inputs and improve translation accuracy even for low resource language pairs (Bojar, para. 14).
Regarding claim 10, Le, as modified above, teaches the method of claim 1.
Le, as modified above, does not teach, however Song teaches
further comprising training the neural network based on each given synthetic sentence pair and the first and second training signals (Song in [0051] teaches In implementations, the predetermined text vector prediction models of the target language and the source language are generated by reading a pre-stored parallel corpus, setting a training goal as to maximize average translation probabilities of sentences in the parallel corpus between the target language and the corresponding source language as background, training a predetermined bilingual encoding and decoding model for text vectors, designating an encoding part of the bilingual encoding and decoding model for text vectors after training as the predetermined text vector prediction model of the source language, and by designating a reverse model of the encoding part of the trained bilingual encoding and decoding model for text vectors as the predetermined text vector prediction model of the target language [here Song is describing training a model based on maximizing probabilities (training signals)]).
Song is considered to be analogous to the claimed invention because it is in the same field of neural networks used for bilingual prediction models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Song to allow for training a model based on maximizing probabilities. Doing so would resolve issues relating to semantic inconsistency of candidate translations. Song, para. 7.
Regarding claim 11, Le, as modified above, teaches the method of claim 1.
Le, as modified above, does not teach, however Bojar teaches
further comprising training the neural network based on a plurality of human-graded sentence pairs (Bojar in [0044] teaches Training a translation model with the corpus FINAL is performed using well known approaches, such as tensor2tensor transformer and RNN—Recurrent neural network architectures. Automatic validation could be then performed using BLEU metric, Meteor, CHRF3 or other suitable automatic metric on both noisy and clean validation corpora. Alternatively as a “translation quality metric score” any scoring algorithm for evaluating the quality of translated text based on existing human translations could be used [existing human translations maps to human graded]).
Bojar is considered to be analogous to the claimed invention because it is in the same field of neural network based training for translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Bojar to allow for comparing to human translations. Doing so would provide a system which is tolerant to noisy inputs and improve translation accuracy even for low resource language pairs (Bojar, para. 14).
Claims 6, 8-9, 17 and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Le, in view of Bojar, in view of Song, in view of Roth et al. (US Patent Pub. No. 2012/0101804), hereinafter Roth.
Regarding claims 6 and 17, Le, as modified above, teaches the method and processing system of claims 5 and 13.
[claim 17 only] wherein the one or more processors are further configured to
Le, as modified above, does not teach, however Roth teaches
wherein generating the plurality of synthetic sentence pairs comprises substituting one or more words of the first passage of text to create the second passage of text (Roth in [0078] teaches The sampling component 45 produces a set of neighbor translations by using a set of operators which are designed to perturb the current target sentence slightly, by making small changes to it, for example by performing one or more of inserting words, removing words, inserting words, reordering the words, and replacing words).
Roth is considered to be analogous to the claimed invention because it is in the same field of natural language processing involving training of translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Roth to allow for substituting one or more words of a text passage. Doing so would allow for generating a plurality of translation neighbors which in turn provides improved training for the translation model (Roth, Abstract).
Regarding claims 8 and 19, Le, as modified above, teaches the method and processing system of claims 1 and 12.
Le, as modified above, does not teach, however Roth teaches
[claim 19 only] wherein the one or more processors are further configured to [substitute]
wherein generating the plurality of graded sentence pairs comprises, for each given graded sentence pair of a first subset of the graded sentence pairs, substituting one or more words of the original passage of text of the given graded sentence pair to create the modified passage of text of the given graded sentence pair (Roth in [0078] teaches The sampling component 45 produces a set of neighbor translations by using a set of operators which are designed to perturb the current target sentence slightly, by making small changes to it, for example by performing one or more of inserting words, removing words, inserting words, reordering the words, and replacing words).
Roth is considered to be analogous to the claimed invention because it is in the same field of natural language processing involving training of translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Roth to allow for replacing one or more words of a text passage. Doing so would allow for generating a plurality of translation neighbors which in turn provides improved training for the translation model (Roth, Abstract).
Regarding claims 9 and 20, Le, as modified above, teaches the method and processing system of claims 8 and 12.
Le, as modified above, does not teach, however Roth teaches
[claim 20 only] wherein the one or more processors are further configured to
wherein generating the plurality of graded sentence pairs includes removing one or more words of the original passage of text to create the modified passage of text (Roth in [0078] teaches The sampling component 45 produces a set of neighbor translations by using a set of operators which are designed to perturb the current target sentence slightly, by making small changes to it, for example by performing one or more of inserting words, removing words, inserting words, reordering the words, and replacing words).
Roth is considered to be analogous to the claimed invention because it is in the same field of natural language processing involving training of translation models. Therefore, it would have been obvious to someone of ordinary skill in the art before the effective filing date of the claimed invention to have modified Le, as modified above, further in view of Roth to allow for removing one or more words of a text passage. Doing so would allow for generating a plurality of translation neighbors which in turn provides improved training for the translation model (Roth, Abstract).
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PAUL J. MUELLER whose telephone number is (571)272-1875. The examiner can normally be reached M-F 9:00am-5:00pm (Eastern).
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Daniel C. Washburn can be reached at 571-272-5551. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
PAUL MUELLER
Examiner
Art Unit 2657
/PAUL J. MUELLER/Examiner, Art Unit 2657