DETAILED ACTION
Notice of 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 .
Response to Arguments
Applicant’s amendments and arguments filed 03/04/2026, with respect to claim(s) 1-8 have been fully considered. Applicant amended claims 1-5, 7, 8, cancelled claim 6 and added new claim 9.
Claim objections for claims 2, 7 have been withdrawn in view of the amended claims filed on 03/04/2026.
35 U.S.C. 112 (f) interpretations of Claims 1-5, 7, 8 have been withdrawn in view of the amended claims filed on 03/04/2026.
Applicant’s arguments in pages 7-9, filed 03/04/2026, with respect to 35 U.S.C 101 rejections of claims 1-5, 7, 8 have been fully considered but they are not persuasive. Applicant argued that the amended claim 1 describes advances in artificial intelligence technologies such as deep learning. The recognition and prediction ability of such artificial intelligence depends on the quantity and quality of training data used to train a model. Therefore, there is a field of technology of generating augmentation data through processing with a plurality of degrees of augmentation for each augmentation method. But throughout the amended claim 1, with the exception of “ processing circuitry” at the beginning of the claim limitations, no specialized or unique technology have been mentioned. The use of a computer does not preclude performance of the invention via pen and paper or in a person’s mind. Also, the use of a computer or other machinery in its ordinary capacity to perform a task or simply adding a general purpose computer to an abstract idea, does not integrate a judicial exception into a practical application. Here the computer is the machine that is merely an object on which the method operates, which does not integrate the exception into a practical application or provide significantly more.
Applicant also argued that the amended features of claim 1, which is incorporated from claim 6, further improve the computer functionality and technological environment while also integrating any abstract idea into a practical application. Examiner respectfully disagrees. Amended feature describes deriving the accuracy of the models in two scenario, once trained with only training data and next trained with augmented data with training data, and based on the results, decide that the model trained with augmented data is the one with highest accuracy. A human can calculate with the pen and paper, the accuracy of a model with two different types of data, from input and the output of the model by using two different data sets and can decide which types of data has the highest accuracy based on the result. There is no details about how or which technology or which unique method is used and how it is improving technology and how it is integrated into a practical application. They are recited as performing generic computer functions routinely used in computer applications, which is not sufficient to amount to significantly more than the judicial exception.
Applicant again argued that in view of Ex Parte Guillaume Desjardins et al., where the panel equated any machine learning with an unpatentable 'algorithm' and the remaining additional elements as 'generic computer components,' without adequate explanation. ... Examiners and panels should not evaluate claims at such a high level of generality and the claims do recite the "improvement" at a proper level of specificity, especially in view of the recent decision of Ex Parte Desjardins. Examiner respectfully disagrees. In Ex parte Desjardins, the claims recite the steps of training a machine learning model by elaborating on first training data, second training data, how the values of plurality of parameters are adjusted to optimize performance of machine learning model. The claim language recites the specialized features on how the training of the model is done and how the improvement is achieved. In contrast, the claims of current application recite the steps of training data augmentation device. The claim language is silent about what type of specialized features are used to perform the task to make improvement to the technology. The claims are not similarly constitute as the claims in Exparte Desjardins. Thus, the 35 U.S.C 101 rejections of Claims 1-5, 7, 8 have been maintained. Newly added claim 9 is a method claim, performing the steps in device claim 1, thus claim 9 is also rejected with 35 U.S.C 101.
Applicant’s arguments filed 03/04/2026, with respect to claim(s) 1-8, under 35 U.S.C.103 have been fully considered but they are not persuasive. Applicant amended claim 1 by incorporating features from claim 6, which was rejected by prior art Mahmud. Applicant argued that Mahmud does not clearly disclose deriving accuracy of each of a plurality of models obtained when (i) the training data and augmentation data of each of a plurality of stages are combined and used for training and (ii) when the training data is exclusively used for training. Examiner respectfully disagrees. Mahmud in para.[0049],[0050] shows that accuracy gain is calculated for different models, base model, augmented model, original model with i) training data ( 1000 examples) combined with 500 examples to augment data and ii) only with training data ( 1000 examples) . Examiner believes, prior reference Mahmud does teach the amended limitations. Thus, the previous 35 U.S.C. rejections of claim 1-5, 7-9 has been maintained. Please see the rejections below.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-5, 7-9 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an
abstract idea without significantly more.
The Independent claim 1, 9 recite “a training data augmentation device comprising: processing circuitry configured to derive accuracy of each of a plurality of models obtained (i) when the training data and augmentation data of each of stages are combined and used for training and (ii) when the training data is exclusively used for training, based on accuracy of an output result obtained by inputting predetermined test data to each of the models”; “and determine, as optimal augmentation data, augmentation data of a stage at which the accuracy of the model is higher than the accuracy when the training data is exclusively used for training and becomes highest accuracy”. The limitations above as drafted, is a process that, under its broadest reasonable interpretation, covers a mental process, as this could be performed in the human mind or with the aid of pen and paper.
The limitation of "generate... ", ‘processing…”, "derive ... ", “determine..”, as drafted covers mental activities. More specifically, a human can process a sentence for augmentation or improvement by replacing words with synonyms or other replacement words, can determine a dependency relationship between word pairs of the each improved sentence and decide which one to use for training purpose, by comparing the results with a predetermined threshold, can calculate the accuracy and determine which one is showing optimal performance. All the steps above are examples of observation and evaluation that could be performed in the human mind or with the aid of pencil and paper.
The claim 1 recites the additional limitation of “processing circuitry”, for performing the method, which is recited at a high level of generality, recited as performing generic computer functions routinely used in computer applications, which is not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
Thus, taken alone, the additional elements do not amount to significantly more than the above identified judicial exception (the abstract idea). Looking at the limitations as an ordered combination adds
nothing that is not already present when looking at the elements taken individually. There is no indication
that the combination of elements improves the functioning of a computer or improves any other technology. Their collective functions merely provide conventional computer implementation. Claims 1 and 9 are therefore not drawn to eligible subject matter as it is directed to an abstract idea without
significantly more than the abstract idea.
Claim 2 recites the additional limitation of “wherein the processing circuitry uses a point-wise mutual information as the degree of association” , where point-wise mutual information as per para.[0021],[0022],[0030], is widely and generally used mathematical information in linguistic data, which could be performed with the aid of pen and paper. The claim recites additional element “processing circuitry” which is recited as performing generic computer functions routinely used in computer applications, and not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
Claim 3 recites “wherein the processing circuitry sets a word pair including: a proper noun; and any one of a noun, an adjective and a verb, each having a dependency relationship with the proper noun, as a target for deriving the degree of association in the word pair having the dependency relationship”, to determine that the word pair has a proper noun and having dependency relationship with a noun, an adjective and a verb, is an evaluation, observation and could be performed in the human mind or with the aid of pen and paper. The claim recites additional element “processing circuitry” which is recited as performing generic computer functions routinely used in computer applications, and not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
Claim 4 recites “wherein, for an augmented sentence including a plurality of word pairs each having a dependency relationship, when there is a word pair whose degree of association is less than or equal to a threshold of a certain stage among the plurality of word pairs, the processing circuitry determines not to add the augmented sentence to augmentation data of the stage”, to determine that the word pairs dependency relationship is less than a threshold and determine not to add it to the augmented sentence, is an evaluation, observation and could be performed in the human mind or with the aid of pen and paper. The claim recites additional element “processing circuitry” which is recited as performing generic computer functions routinely used in computer applications, and not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
Claim 5 recites “wherein, for an augmented sentence including a plurality of word pairs each having a dependency relationship, when all degrees of association in the plurality of word pairs are less than or equal to a threshold of a certain stage, the processing circuitry determines not to add the augmented sentence to augmentation data of the stage”, to determine that the word pairs dependency relationship is less than a threshold or zero and determine not to add it to the augmented sentence, is an evaluation, observation and could be performed in the human mind or with the aid of pen and paper. The claim recites additional element “processing circuitry” which is recited as performing generic computer functions routinely used in computer applications, and not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
Claim 7 recites “wherein the processing circuitry derives the accuracy of the model, on a further basis of accuracy of an output result obtained by inputting predetermined test data to a model obtained when the training data and at least two pieces of augmentation data of each of the stages are combined and used for training”, where the accuracy of a model can be determined on the basis of accuracy of an output result, could be performed with the aid of pen and paper. The claim recites additional element “processing circuitry”, which is recited as performing generic computer functions routinely used in computer applications, and not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
Claim 8 recites “wherein the processing circuitry sets a word pair including: a proper noun; and any one of a noun, an adjective and a verb, each having a dependency relationship with the proper noun, as a target for deriving the degree of association in the word pair having the dependency relationship”, to determine that the word pair has a proper noun and having dependency relationship with a noun, an adjective and a verb, is an evaluation, observation and could be performed in the human mind or with the aid of pen and paper. The claim recites additional element “processing circuitry” which is recited as performing generic computer functions routinely used in computer applications, and not sufficient to amount to significantly more than the judicial exception. The claim as drafted, is not patent eligible.
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.
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 of this title, 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, 2, 4, 5, 7 and 9 are rejected under 35 U.S.C. 103 as being unpatentable over Goutal et al. ( US 20200159993 A1), hereinafter referenced as Goutal, in view of Abhishek et al. (US 20220366293 A1), hereinafter referenced as Abhishek, further in view of Mahmud et al. (US 20200372395 A1), hereinafter referenced as Mahmud.
Regarding Claim 1, Goutal teaches a training data augmentation device comprising:
processing circuitry configured to ( Goutal: Para.[0036], Fig. 2 illustrates a data augmentation method, where at block B204, original text is then input to another transformation that replaces at least some of the words of the electronic text document presented at its input with synonyms);
[[and]] sentence determined to be added ( Goutal: Para.[0035], [0036], Fig. 2, TextDataAugmentationFunction takes a text document OriginalText as an input, and outputs a text document AugmentedText. This function applies successive transformations to OriginalText to produce AugmentedText, as shown at B202, B204. At B208, a similarity measure, based on OriginalText, is computed for the resultant successively-transformed AugmentedText. If, as shown at B210, the similarity measure SimMeasure is greater or equal to a similarity measure threshold SimMeasureThreshold, then the augmented electronic text document is kept as shown at B212. If, SimMeasure is less than the similarity measure threshold SimMeasureThreshold, the successively-transformed AugmentedText may be discarded, as shown at B214, and may not be used for further training).
Goutal while teaching the device of claim 1, fails to explicitly teach the claimed, to derive a degree of association in a word pair having a dependency relationship; derive accuracy of each of a plurality of models obtained (i) when the training data and augmentation data of each of stages are combined and used for training and (ii) when the training data is exclusively used for training, based on accuracy of an output result obtained by inputting predetermined test data to each of the models”; “and determine, as optimal augmentation data, augmentation data of a stage at which the accuracy of the model is higher than the accuracy when the training data is exclusively used for training and becomes highest accuracy
However, Abhishek does teach the claimed, analyzing, to derive a degree of association in a word pair having a dependency relationship ( Abhishek: Para.[0031], [0034], Fig. 2 illustrates a block diagram of a system for determining words in a text record that can be replaced. Dependency module 206 uses a dependency parser model to generate respective dependency scores for the words in the text record. The dependency parser model determines a syntactic dependency relationship between the words) ;
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Abhishek’s teaching of augmenting text of a small class for text classification, into the system and method of method of generating an augmented electronic text document, taught by Goutal, because, the approach of text augmentation of a small class to balance a dataset would improve the performance of a text classification machine learning model in terms of accuracy and generalizability.(Abhishek, Para.[0021]).
Goutal in view of Abhishek while teaching the device of claim 1, fail to explicitly teach the claimed, derive accuracy of each of a plurality of models obtained (i) when the training data and augmentation data of each of stages are combined and used for training and (ii) when the training data is exclusively used for training, based on accuracy of an output result obtained by inputting predetermined test data to each of the models; and determine, as optimal augmentation data, augmentation data of a stage at which the accuracy of the model is higher than the accuracy when the training data is exclusively used for training and becomes highest accuracy.
However, Mahmud does teach the claimed, further comprising: derive accuracy of each of a plurality of models obtained (i) when the training data and augmentation data of each of stages are combined and used for training and (ii) when the training data is exclusively used for training, based on accuracy of an output result obtained by inputting predetermined test data to each of the models ( Mahmud: Para.[0049], Accuracy gain from base model on an original test set is the difference in accuracy for an original test set when an original model and an augmented model are applied. The augmented model for the original test set is used to compute an accuracy score based on the augmented model's prediction on that test set. The same is done for the original model, and the difference is the accuracy gain);
and determine, as optimal augmentation data, augmentation data of a stage at which the accuracy of the model is higher than the accuracy when the training data is exclusively used for training and becomes highest accuracy ( Mahmud: Para.[0050], if the original test set has one thousand examples and they are augmented by another five hundred examples, then the augmented test set has fifteen hundred examples. The accuracy of the original model on the augmented test set might be 72% and the accuracy of the augmented model on the augmented test set might be 75%, so the accuracy gain is 3%).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Mahmud’s teaching of data augmentation for text based AI applications, into the system and method, taught by Goutal in view of Abhishek, because, by applying a set of augmentation approaches with different parameters, controlling the validation process, the best augmented model for a particular cognitive system can be selected. (Mahmud, Para.[0022]).
Claim 9 is a method claim performing the steps in device claim 1 above and as such, claim 9 is similar in scope and content to claim 1 and therefore, claim 9 is rejected under similar rationale as presented against claim 1 above.
Regarding Claim 2, Goutal in view of Abhishek, further in view of Mahmud teach the training data augmentation device according to claim 1. Abhishek further teaches, wherein the processing circuitry uses a point-wise mutual information as the degree of association ( Abhishek: Para.[0026], Fig. 1, text augmentation module 108 sends each word to suitable word generation module 112, which generates a suitable word list for a given word based on suitable word scores that are based on cosine similarity scores with class-based word importance statistics, contextual probability scores with class-based word importance statistics, and synonym-based class word scores).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Abhishek’s teaching of augmenting text of a small class for text classification, into the system and method, taught by Goutal in view of Mahmud, because, the approach of text augmentation of a small class to balance a dataset would improve the performance of a text classification machine learning model in terms of accuracy and generalizability.(Abhishek, Para.[0021]).
Regarding Claim 4, Goutal in view of Abhishek , further in view of Mahmud teach the training data augmentation device according to claim 1 . Goutal further teaches, wherein, for an augmented sentence including a plurality of word pairs each having a dependency relationship, when there is a word pair whose degree of association is less than or equal to a threshold of a certain stage among the plurality of word pairs, the processing circuitry determines not to add the augmented sentence to augmentation data of the stage ( Goutal: Para.[0035], [0036], Fig. 2, TextDataAugmentationFunction takes a text document Original Text as an input, and outputs a text document Augmented Text. At B208, a similarity measure, based on OriginalText, is computed for the resultant successively-transformed AugmentedText. If, SimMeasure is less than the similarity measure threshold SimMeasureThreshold, the successively-transformed AugmentedText may be discarded, as shown at B214, and may not be used for further training).
Regarding Claim 5, Goutal in view of Abhishek, further in view of Mahmud teach the training data augmentation device according to claim 1. Goutal further teaches, wherein, for an augmented sentence including a plurality of word pairs each having a dependency relationship, when all degrees of association in the plurality of word pairs are less than or equal to a threshold of a certain stage, the processing circuitry determines not to add the augmented sentence to augmentation data of the stage ( Goutal: Para.[0035], [0036], Fig. 2, TextDataAugmentationFunction takes a text document Original Text as an input, and outputs a text document Augmented Text. At B208, a similarity measure, based on OriginalText, is computed for the resultant successively-transformed AugmentedText. If, SimMeasure is less than the similarity measure threshold SimMeasureThreshold, the successively-transformed AugmentedText may be discarded, as shown at B214, and may not be used for further training).
Regarding Claim 7, Goutal in view of Abhishek, further in view of Mahmud teach the training data augmentation device according to claim [[6]] 1. Mahmud further teaches, wherein theprocessing circuitry derives the accuracy of the model, on a further basis of accuracy of an output result obtained by inputting predetermined test data to a model obtained when the training data and at least two pieces of augmentation data of each of the stages are combined and used for training ( Mahmud: Para.[0054], Fig. 5, after selection of the optimum augmentation method, it can be used to construct a more comprehensive training data set for the cognitive system. The selected data augmentation method 92 is applied to a training set 94. The result is an augmented training data set 96 which is then used to train the cognitive system, yielding an optimized deep question/answer 98).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Mahmud’s teaching of data augmentation for text based AI applications, into the system and method, taught by Goutal in view of Abhishek, because, by applying a set of augmentation approaches with different parameters, controlling the validation process, the best augmented model for a particular cognitive system can be selected. (Mahmud, Para.[0022]).
Claims 3 and 8 are rejected under 35 U.S.C. 103 as being unpatentable over Goutal et al. ( US 20200159993 A1), hereinafter referenced as Goutal, in view of Abhishek et al. (US 20220366293 A1), hereinafter referenced as Abhishek, further in view of Mahmud et al. (US 20200372395 A1), hereinafter referenced as Mahmud , further in view of Zhu et al. (US 20200320113 A1), hereinafter referenced as Zhu.
Regarding Claim 3, Goutal in view of Abhishek, further in view of Mahmud teach the training data augmentation device according to claim 1. Goutal in view of Abhishek further in view of Mahmud fail to explicitly teach the claimed, wherein the processing circuitry sets a word pair including: a proper noun; and any one of a noun, an adjective and a verb, each having a dependency relationship with the proper noun, as a target for deriving the degree of association in the word pair having the dependency relationship.
However, Zhu does teach the claimed, wherein the processing circuitry sets a word pair including:
a proper noun ( Zhu: Para.[0039], word pair “Nike” and “Shoes”, where “Nike” is Proper noun) ;
and any one of a noun, an adjective and a verb, each having a dependency relationship with the proper noun ( Zhu: Para.[0038], [0039], word pair “Nike” and “Shoes”, where “Nike” is Proper noun, “ Shoes” is a noun having dependency relationship. Also, in “Nike running shoe”, verb running has a dependency relationship with Nike),
as a target for deriving the degree of association in the word pair having the dependency relationship ( Zhu: Para.[0027],[0039], word pairs are evaluated to determine syntactic information such as dependency (how words depend on each other) and co-reference information).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Zhu’s teaching of method and system for syntactic searching, into the system and method, taught by Goutal in view of Abhishek, further in view of Mahmud, because, by comparing each word of a search term with previously stored database results such that more relevant search results are retrieved and displayed to the user, can improve searching functionality. (Zhu, Para.[0007]-[0010]).
Claim 8 is a device claim performing the steps in device claim 3 above and as such, claim 8 is similar in scope and content to claim 3 and therefore, claim 8 is rejected under similar rationale as presented against claim 3 above.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NADIRA SULTANA whose telephone number is (571)272-4048. The examiner can normally be reached M-F,7:30 am-5:00pm.
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, Paras D. Shah can be reached on (571) 270-1650. 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.
/NADIRA SULTANA/Examiner, Art Unit 2653
/Paras D Shah/Supervisory Patent Examiner, Art Unit 2653
05/28/2026
Prosecution Insights