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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 3/14/25 and 7/20/26 and is being considered by the examiner.
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 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.
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 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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-3, 5-7, 9, 12, 19-20, 24-29 are rejected under 35 U.S.C. 103 as being unpatentable over Dvorkovich (US 20220198159 A1) in further view of Lee (US 11093719 B2)
(Claim 1) (Original) A method for training a translation model
(Claim 19) (Currently Amended) An electronic device, comprising: a memory; and a processor coupled to the memory, wherein the memory has, stored therein, executable instructions that, when executed by the processor, cause the electronic device to perform ([0065] Returning to the description of FIG. 1, the system 100 also comprises the server 112 that can be implemented as a conventional computer server. In the depicted non-limiting embodiments of the present technology, the server 112 is a single server. In alternative non-limiting embodiments of the present technology, functionalities of the server 112 may be distributed and may be implemented via multiple servers. The server 112 may include one or more processors, one or more non-transitory memory devices, computer-readable instructions, and/or additional hardware components, additional software components, and/or combination thereof, for implementing various functionalities of the server 112, without departing from the scope of the present technology.)
(Claim 20) (Currently Amended) A non-transitory computer-readable storage medium having, stored thereon, executable instructions that, when executed by a processor, implement a method ([0065] Returning to the description of FIG. 1, the system 100 also comprises the server 112 that can be implemented as a conventional computer server. In the depicted non-limiting embodiments of the present technology, the server 112 is a single server. In alternative non-limiting embodiments of the present technology, functionalities of the server 112 may be distributed and may be implemented via multiple servers. The server 112 may include one or more processors, one or more non-transitory memory devices, computer-readable instructions, and/or additional hardware components, additional software components, and/or combination thereof, for implementing various functionalities of the server 112, without departing from the scope of the present technology.)
applying second type of sample data to a back-translation model associated with the translation model to obtain training sample data(Dvorkovich ¶[0103] The server 112 may be configured to use the second translation model 130 in the direction 420 for generating artificial examples of translation between the rare language and the target language. To that end, the server 112 may be configured to retrieve the plurality of clean sentences 390 in the target language (e.g., accurately written sentences in Russian) [target language] and input them into the second translation model 130 for performing back translation into Chuvash. [back-translation] As such, the server 112 may be configured to generate an artificial sentence 451 [training sample data] for the sentence 391, an artificial sentence 452 for the sentence 392, and an artificial sentence)
training the translation model based on the training sample data (Dvorkovich ¶[0108] The server 112 may be configured to use the plurality of artificial examples 480 [training sample data] in addition to the plurality of actual examples 310 for training [training] the first translation model 120. As it will now be described with reference to FIG. 5, the server 112 may be configured to use the transliteration model 140 in order to generate a plurality synthetic actual examples based on the plurality of actual examples 310 and a plurality of synthetic artificial examples based on the plurality of artificial examples 480).
Dvorkovich does not explicitly disclose however Lee teaches the translation model being capable of converting first type of data into second type of data, the method comprising (Lee ¶Col6ll34-45 The machine translation apparatus receives, as an input, [first type of data=speech] the source sentence as text or a speech signal from a user, ¶Col6ll56-62 In one example, when the back-translation is used, a target sentence B generated by translating a source sentence A is translated back [back-translation] into the source language to determine a comparison sentence A′, and a reliability of the target sentence B is determined based on a similarity between the source sentence A and the comparison sentence A′., ¶(Col2ll42-48 The processor-implemented machine translation method may further include outputting [second type=text] either the target sentence or the re-determined target sentence as either one or both of text and a speech signal in response to a determination that either the target sentence or the re-determined target sentence is appropriate as the translation result.)
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify backtranslation of Dvorkovich to include data types of Lee in order to enhance flexibility and provide real-time accessibility.
With respect to claim 2 Lee further teaches wherein the first type is one of a speech type and a text type, and the second type is the other of the speech type and the text type(Lee ¶Col6ll34-45 The machine translation apparatus receives, as an input, the source [first type of data=speech] sentence as text or a speech signal from a user, ¶Col6ll56-62 In one example, when the back-translation is used, a target sentence B generated by translating a source sentence A is translated back [back-translation] into the source language to determine a comparison sentence A′, and a reliability of the target sentence B is determined based on a similarity between the source sentence A and the comparison sentence A′., ¶(Col2ll42-48 The processor-implemented machine translation method may further include outputting [second type=text] [second type=text] either the target sentence or the re-determined target sentence as either one or both of text and a speech signal in response to a determination that either the target sentence or the re-determined target sentence is appropriate as the translation result.).
With respect to claim(s) 3 Lee further teaches wherein the first type of data and the training sample type are both speech data in discrete form, or wherein the first type of data is a continuous speech signal, and the training sample data is speech data in discrete form (Lee ¶Col6ll34-45 The machine translation apparatus receives, as an input, the source [first type of data=speech] sentence as text or a speech signal from a user, ¶Col6ll56-62 In one example, when the back-translation is used, a target sentence B generated by translating a source sentence A is translated back [back-translation] into the source language to determine a comparison sentence A′, and a reliability of the target sentence B is determined based on a similarity between the source sentence A and the comparison sentence A′., ¶(Col2ll42-48 The processor-implemented machine translation method may further include outputting [second type=text] [second type=text] either the target sentence or the re-determined target sentence as either one or both of text and a speech signal in response to a determination that either the target sentence or the re-determined target sentence is appropriate as the translation result.)
With respect to claim(s) 5, 24 and 27 Dvorkovich wherein the back-translation model comprises a model that is backward-constructed and matched with the translation model (Dvorkovich ¶[0103] The server 112 may be configured to use the second translation model 130 in the direction 420 [matched with the translation model] for generating artificial examples of translation between the rare language and the target language. To that end, the server 112 may be configured to retrieve the plurality of clean sentences 390 in the target language (e.g., accurately written sentences in Russian) and input them into the second translation model 130 for performing back translation into Chuvash. [back-translation] As such, the server 112 may be configured to generate an artificial sentence 451 [training sample data] for the sentence 391, an artificial sentence 452 for the sentence 392, and an artificial sentence.)
With respect to claim(s) 6, 25 and 28 Dvorkovich teaches wherein applying the second type of sample data to the back-translation model associated with the translation model to obtain the training sample data, comprises: using the second type of sample data as an input to the back-translation model, to obtain model output data; and the model output data or data obtained after processing the model output data are used as the training sample data, wherein the processing of the model output data comprises processing that generates data perturbation (Dvorkovich ¶[0103] The server 112 may be configured to use the second translation model 130 in the direction 420 for generating artificial examples of translation between the rare language and the target language. To that end, the server 112 may be configured to retrieve the plurality of clean sentences 390 [second type] in the target language (e.g., accurately written sentences in Russian) and input them into the second translation model 130 for performing back [back-translation] translation into Chuvash. [model output] As such, the server 112 may be configured to generate an artificial sentence 451 for the sentence 391, an artificial sentence 452 for the sentence 392, and an artificial sentence.) Examiner Note: based on the “or” the examiner is not mapping the last limitation regarding perturbation.
With respect to claim(s) 7, 26 and 29 Dvorkovich teaches wherein the first type of data comprises a continuous speech signal, and the translation model comprises (Lee ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. [first type of data is discretized to a sentence] The machine translation apparatus determines the target sentence from the source sentence based [second type of data is discretized as a target sentence] on a neural network-based translation model that translates the source language into the target language):
Lee further teaches a discretization module configured to extract speech data in discrete form from the continuous speech signal(Lee ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. [first type of data is discretized to a sentence] The machine translation apparatus determines the target sentence from the source sentence based [second type of data is discretized as a target sentence] on a neural network-based translation model that translates the source language into the target language); and
a translation module configured to convert the speech data in discrete form into a text in a target language as the second type of data (Lee ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. [first type of data is discretized to a sentence] The machine translation apparatus determines the target sentence from the source sentence based [second type of data is discretized as a target sentence] on a neural network-based translation model that translates the source language into the target language)
With respect to claim 9, Dvorkovich teaches wherein the back-translation model comprises a model that is backward-constructed and matched with the machine translation module, and the model is capable of obtaining, from training texts in the target language as the second type of sample data, specific text data as the training sample data (Dvorkovich ¶[0103] The server 112 may be configured to use the second translation model 130 in the direction 420 for generating artificial examples of translation between the rare language and the target language. To that end, the server 112 may be configured to retrieve the plurality of clean sentences 390 in the target language (e.g., accurately written sentences in Russian) [Target language] and input them into the second translation model 130 for performing back translation into Chuvash. [back-translation] As such, the server 112 may be configured to generate an artificial sentence 451 [training sample data] for the sentence 391, an artificial sentence 452 for the sentence 392, and an artificial sentence)
Lee further teaches wherein the first type of data comprises a continuous speech signal (Lee ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. [first type of data is discretized to a sentence] The machine translation apparatus determines the target sentence from the source sentence based [second type of data is discretized as a target sentence] on a neural network-based translation model that translates the source language into the target language):
and the translation model comprises: a discretization module configured to extract speech data in discrete form from the continuous speech signal (Lee ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. [first type of data is discretized to a sentence] The machine translation apparatus determines the target sentence from the source sentence based [second type of data is discretized as a target sentence] on a neural network-based translation model that translates the source language into the target language):
a speech conversion module configured to convert the speech data in discrete form into intermediate text data; (Lee ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. [first type of data is discretized to a sentence] The machine translation apparatus determines the target sentence from the source sentence based [second type of data is discretized as a target sentence] on a neural network-based translation model that translates the source language into the target language);
and a machine translation module configured to convert the intermediate text data into a text in a target language as the second type of data ( ¶Col6 ll34-45 For example, when the source sentence is input as a speech signal, a speech recognition process of converting a speech signal to a source sentence is additionally performed. The machine [machine translation] translation apparatus determines the target sentence from the source sentence based on a neural network-based translation model that translates the source language into the target language).
With respect to claim 12 Dvorkovich teaches wherein the training the translation model based on the training sample data comprises: training the translation model with the training sample data as an input and the second type of sample data as an output (Dvorkovich ¶[0108] The server 112 may be configured to use the plurality of artificial examples 480 in addition to the plurality of actual examples 310 for training the first translation model 120, ¶[0113] The server 112 may make use of the synthetic actual example 560 for generating a synthetic actual training set 610. The server 112 may be configured to input the synthetic sentence 502 into the first translation model 120 and the first translation model 120 is configured to output a sentence 612 [second type] in the target language.)
Claims 11 and 23 are rejected under 35 U.S.C. 103 as being unpatentable over Dvorkovich, Lee in further view of Ramachandrula (US 20160098622 A1).
With respect to claim 11, none of Dvorkovich and Lee explicitly disclose however Ramachandrula teaches wherein the discretization module is configured to extract the speech data in discrete form from the continuous speech signal based on a vector quantization method and/or a clustering method (Ramachandrula ¶[0006] FIG. 4 illustrates a feature vector extraction and vector quantization [vector quantization] encoding of a speech signal, according to an example.)
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify backtranslation of Dvorkovich in view of data types of Lee to include quantization of Ramachandrula in order to provide noise reduction and significant data compression.
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Dvorkovich, Lee in further view of Jiang (US 20160098622 A1).
With respect to claim 13, none of Dvorkovich and Lee explicitly disclose however Jiang teaches performing training of the back-translation model in parallel with the training of the translation model (Jiang ¶[0008] when a sum of the forward translation similarity and the reverse translation similarity converges, it is determined that training of the bidirectional translation model is completed.)
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify backtranslation of Dvorkovich in view of data types of Lee to include parallel training of Jiang in order to improve translation accuracy through iterative refinement.
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Dvorkovich, Lee in further view of Zhang (US 20210312902 A1).
With respect to claim 14, none of Dvorkovich and Lee explicitly disclose however Zhang teaches wherein the obtaining the training sample data and the model training are iteratively performed, until a specific iteration termination condition is satisfied (Zhang ¶[0049] The training sample is iteratively processed by steps S302-S305 until a second loss function is minimized).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify backtranslation of Dvorkovich in view of data types of Lee to include iterative training of Zhang in order to improve data efficiency and reduce labelling costs.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ATHAR N PASHA whose telephone number is (408)918-7675. The examiner can normally be reached on Monday-Thursday Alternate Fridays, 7:30-4:30 PT.
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 Washburn can be reached on (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 an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see https://ppair-my.uspto.gov/pair/PrivatePair. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/ATHAR N PASHA/Primary Examiner, Art Unit 2657
Prosecution Insights