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 .
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.
Step 1 analysis for all claims:
Claims 1-14 and 16-20 are directed to a method (process) Therefore these claims fall within one of the statutory categories of invention (process, manufacture, composition of matter or machine). As such, the claim is eligible under 35 U.S.C. 101.
Claim 15 is rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim recites “A computer-readable storage medium”, and the specification fails to properly describe this computer-readable storage medium as being non-transitory. Therefore, it is not clear what exactly the claimed computer readable storage medium is, as it cannot be clearly concluded that this medium is being described as excluding carrier waves and/or transient signals. Therefore, based on the broadest reasonable interpretation, it can by any type of storage, and this encompasses signals or transmission media which are not statutory embodiments under 35 USC 101. As such, it fails to fall within a statutory category under 35 USC 101.
Claim 1:
Step 2A, Prong 1 analysis:
As discussed above, the additional elements of “obtaining, from a user, original data for the target deep learning model” which is recited at a high level of generality and amounts to extra-solution activity of receiving data i.e. pre-solution activity of gathering data for use in the claimed process. The courts have found limitations directed to obtaining information electronically, recited at a high level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
“determining a type of the original data wherein the type of the original data comprises categorical data with label, session data with label, and data without label, a label of the categorical data indicates a category of the categorical data, and a label of the session data indicates a question-answer relevance of the session data;”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses identifying a data’s type. This would be seen as a mental process because a person having ordinary skill of the art would be able to determine a data’s type by viewing it i.e. as an integer, floating point value, Boolean value, etc.
“and generating the training data according to the type of the original data.;”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses generating data that corresponds to the original data, to be used to train a model. This would be seen as a mental process because a person having ordinary skill of the art would be able to generate data based of the original data, and use that data to train a learning model.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Claim 2:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “generating, in response to the original data being the categorical data, the training data according to the category indicated by the label of the categorical data.”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses creating training data based on a category decided by a label. This would be seen as a mental process because a person having ordinary skill of the art would be able to determine the data’s category, and create training data that adheres to said category.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Claim 3:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “selecting part or all of the categorical data as reference samples;”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses selecting one or multiple categorically sorted pieces of data to utilize as reference samples. This would be seen as a mental process because a person having ordinary skill of the art would be able to distinctly select data that would best fit as a reference sample.
• “using each of the reference samples as a target reference sample;” As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses selecting one or multiple categorically sorted pieces of data to utilize as reference samples. This would be seen as a mental process because a person having ordinary skill of the art would be able to distinctly select data that would best fit as a reference sample.
• “determining the categorical data in a same category as the target reference sample as a positive sample associated with the target reference sample;”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses identifying the data in the same category as the target sample to be used as a positive sample. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider the data that is in the category associated with the target reference sample and using the data as a positive sample.
• “determining the categorical data in a category different from the target reference sample as a negative sample associated with the target reference sample;” As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses identifying the data in a different category as the target sample to be used as a negative sample. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider the data that is in an opposing category that is not associated with the target reference sample and using the data as a negative sample.
• “and grouping the target reference sample, the positive sample associated with the target reference sample and the negative sample associated with the target reference sample into a group of training data.;” As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation grouping three different types of data into one dataset to be used as training data. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider the positive, negative, and reference data samples and group them together to further be used as training data.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Claim 5:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “wherein the step of generating the training data according to the type of the original data comprises: generating, in response to the original data being the session data, the training data according to a question-answer relevance indicated by the label of the session data.”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses generating data based off of a question-and-answer system that is indicated by the label given to the session data. This would be seen as a mental process because a person having ordinary skill of the art since they would be able to observe the question-and-answer relevance and generate training data that corresponds accordingly.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception
Claim 6:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “determining the first matching samples as a positive sample, wherein labels of the first matching samples indicate positive question answer relevance”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses identifying the first matching samples as having a positive question-answer relevance. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider the first matching samples and determine that they have a positive question-answer relevance.
• “determining second matching samples as negative samples, wherein labels of the second matching samples indicate negative question answer relevance”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses identifying the second matching samples as having a negative question-answer relevance. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider the second matching samples and determine that they have a negative question-answer relevance.
• “and grouping the reference sample, the positive samples and the negative samples into a group of training data.”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses grouping the three types of samples into one dataset to be used as training data. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider all three sample types and group them all into one set to be used as training data.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Claim 7:
Step 2A, Prong 2 analysis:
The claim is directed to the same abstract ideas as identified above.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 8:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “generating, in response to the original data being the data without label, the training data by using data enhancement techniques”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses utilizing data enhancement techniques to generate training data. This would be seen as a mental process because a person having ordinary skill of the art since they would be able to utilize the mentioned data enhancing techniques to generate training data.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 9:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “generating a plurality of positive samples from the reference sample by using the data enhancement techniques”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses generating specifically positive samples via the reference sample, by utilizing data enhancement techniques. This would be seen as a mental process because a person having ordinary skill of the art would be able to generate positive samples after considering the reference samples, and utilizing data enhancement techniques.
• “ using each piece of the data without label as a reference sample;”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses using data that is without a label as a reference sample. This would be seen as a mental process because a person having ordinary skill of the art would be able to determine which pieces of data are without labels, and use them as reference samples.
• “generating a plurality of negative samples from the data without label other than the reference sample by using the data enhancement techniques.” As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses generating specifically negative samples via the reference sample, by utilizing data enhancement techniques. This would be seen as a mental process because a person having ordinary skill of the art would be able to generate negative samples after considering the reference samples, and utilizing data enhancement techniques.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 10:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “wherein the data without label is a picture, and the data enhancement techniques comprise performing least one of comprising flipping, mirroring, and cropping on the picture.”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses manipulating an image. This would be seen as a mental process because a person having ordinary skill of the art would be able to obtain the picture and perform one or all of the mentioned data manipulation techniques.
Step 2A, Prong 2 analysis:
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B analysis:
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 11:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “wherein the data without label is a text, and the data enhancement techniques comprise performing a random mask operation on the text.”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses performing a random masking technique on a string of text. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider a piece of text and either manually or recursively apply a random masking operation.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Claim 12:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
• “wherein the data without label is a sound segment, and the data enhancement techniques comprise performing a random mask operation on the sound segment.”. As drafted and under its broadest reasonable interpretation, this limitation covers performance of the limitation in the mind (including an observation, evaluation, judgment, opinion) or with the aid of pencil and paper but for the recitation of generic computer components. For example, this limitation encompasses performing a random masking technique on a sound segment. This would be seen as a mental process because a person having ordinary skill of the art would be able to consider a sound segment and either manually or recursively apply a random masking operation.
Step 2A, Prong 2 analysis:
There are no additional elements that individually or in combination integrate the judicial element into practical application.
Step 2B analysis:
There are no additional elements that individually or in combination amount to significantly more than the judicial exception.
Claim 13:
Claim 13 recites similarly to claim 1, therefore it is rejected under the same basis
Claim 14:
Step 2A, Prong 1 analysis:
The claim is directed to the same abstract ideas as identified above.
Step 2A, Prong 2 analysis:
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
• “at least one processor” which is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicial exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. See MPEP 2106.05(g).
• “at least one memory storing a computer program” which is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicia I exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B analysis:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
• “at least one processor” This limitation is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicia I exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
• “at least one memory storing a computer program” This limitation is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicia I exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 15 as explained above, fails to fall within a statutory category under 35 USC 101, however, this claim can be amended to fall within a statutory category if it was to be directed towards a non-transitory computer-readable storage medium (as an example) instead of a computer readable storage medium; therefore, Examiner will proceed to determine whether such an amended claim would qualify as eligible under Steps 2A and 2B per MPEP 2106 III: “Summary Of Analysis And Flowchart”.
Claim 15:
Step 2A, Prong 1 analysis:
The claim(s) recite(s) in part:
The claim is directed to the same abstract ideas as identified above.
Step 2A, Prong 2 analysis:
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of:
• “A computer-readable storage medium” which is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicial exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. See MPEP 2106.05(g).
• “storing a computer program” which is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicia I exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B analysis:
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional elements of:
• “A computer-readable storage medium” This limitation is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicia I exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
• “storing a computer program” This limitation is recited at a high level of generality and amount to adding the words "apply it" (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea. See MPEP 2106.05(f). As explained by the Supreme Court; in order to make a claim directed to a judicia I exception patent-eligible, the additional element or combination of elements must do '"more than simply state the judicial exception] while adding the words 'apply it"'. Alice Corp. v. CLS Bank, 573 U.S. 208, 221, 110 USPQ2d 1976, 1982-83 (2014) (quoting Mayo Collaborative Servs. V. Prometheus Labs., Inc., 566 U.S. 66, 72, 101 USPQ2d 1961, 1965). Thus, limitations that amount to nothing more than an instruction to apply the abstract idea using a generic computer do not amount to significantly more than the exception itself and do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea.
Accordingly, at Step 2B the additional elements individually or in combination do not amount to significantly more than the judicial exception.
Claim 16:
Claim 16 is recited similarly to claim 4, therefore it is rejected under the same basis.
Claim 17:
Claim 17 is recited similarly to claim 7, therefore it is rejected under the same basis.
Claim 18:
Claim 18 is recited similarly to claim 10, therefore it is rejected under the same basis.
Claim 19:
Claim 19 is recited similarly to claim 11, therefore it is rejected under the same basis.
Claim 20:
Claim 20 is recited similarly to claim 12, therefore it is rejected under the same basis.
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.
Regarding Claims 1, 2, 4, 5, 6, 13, 16
Claims 1 and 13 are rejected under 35 U.S.C. 103 as being unpatentable over Lixia et al., (CN112069302B, referred to as Lixia hereinafter), in view of Kumaran et al., (US10789944B2, referred to as Kumaran hereinafter).
Lixia teaches:
A method for generating training data, the training data being used- configured for training a target deep learning model, the method comprising: (Lixia)[pg.2, pp. 0008]” This application provides a training method, a conversation intent recognition method, and an apparatus for a conversation intent recognition model.” Lixia teaches a method of training a deep learning model using a conversational approach, similar to the question-answer relevance method of training a deep learning model as taught by the claim. Under the broadest reasonable interpretation, it would have been obvious to a person having ordinary skill of the art to use a method of training a learning model that converses with the model via questions and answers.
wherein the type of the original data comprises categorical data with label, (Lixia)[pg.16 pp. 0089]”the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches using classification tags to categorize types of emotions, the tag data mentioned by Lixia being the labels, and the types being the categorical data, as taught by the claim. Based on the broadest reasonable interpretation, the method of using labels to categorize or group different types of the same data as mentioned, would be obvious to a person having ordinary skill of the art.
a label of the categorical data indicates a category of the categorical data, (Lixia)[pg.16 pp. 0089]”the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches using classification tags to categorize types of emotions. Based on the broadest reasonable interpretation, the method of using labels to categorize or group different types of the same data as mentioned, would be obvious to a person having ordinary skill of the art.
obtaining, from a user, original data for the target deep learning model; (Lixia) “ when receiving a conversation input by a user, “ Lixia teaches receiving input from a user that will then be used to train the learning model, as taught by the claim.
determining a type of the original data (Lixia) “when receiving a conversation input by a user,” It is obvious that during the process of receiving an input from a user, the type of the data would have to be determined before it can be used, as taught by the claim.
session data with label, and data without label (Lixia)[pg.16 pp. 0089]”the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” It is obvious that since there is data that is labeled, there must also be data that goes unlabeled, teaching data with and without label, as taught by the claim.
However, Kumaran teaches:
and a label of the session data indicates a question-answer relevance of the session data; (Kumaran)[Abstract] “The processor is configured to generate a machine learning model based on the question, the positive label assigned to the semantically relevant answer,” Kumaran teaches of assigning a positive value to an answer to indicate it’s correctness. Under the claims broadest reasonable interpretation, this limitation would be obvious to person having ordinary skill of the art since both assign labels that indicate if an answer is correct or incorrect.
and generating the training data according to the type of the original data. (Kumaran)[col. 7, lines 41-43]“For each indexed question, the training module 310 generates and trains the machine learning model 312 by providing as input to the machine learning model 312” Kumaran teaches of training a machine learning model using each indexed question. Although not explicitly described, it would be obvious to a person having ordinary skill of the art that the training data is generated according to the original data that is used to generate it. Under the claims broadest reasonable interpretation, this limitation would be obvious to person having ordinary skill of the art since both teach generating training data based on the original data that is used to generate the training data.
It would be obvious to a person having ordinary skill of the art to combine the conversation-intention recognition model training method as taught by Lixia, with the indication of a semantically relevant answer as taught by Kumaran. A person having ordinary skill of the art would be motivated to do so that the model can learn new meanings of new knowledge and original knowledge more accurately (Lixia)[Abstract], and to provide the machine learning model to facilitate a determination of whether a semantically relevant answer can be provided with respect to a subsequent question (Kumaran) [Abstract].
Regarding claim 2:
Lixia teaches:
generating, in response to the original data being the categorical data, the training data according to the category indicated by the label of the categorical data. (Lixia) [pg.16 pp. 0089]” the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches using classification tags to categorize types of emotions. Based on the broadest reasonable interpretation, the method of using labels to categorize or group different types of the same data as mentioned, would be obvious to a person having ordinary skill of the art.
Regarding claim 4:
Lixia teaches:
wherein the categorical data comprises a plurality of labels, and the category of the categorical data is determined by at least one of the plurality of labels of the categorical data (Lixia) [pg.16 pp. 0089]” the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches using a plurality of classification tags to categorize types of emotions. Based on the broadest reasonable interpretation, the method of using labels to categorize or group different types of the same data as mentioned, would be obvious to a person having ordinary skill of the art.
Regarding claim 5:
Kumaran teaches:
generating, in response to the original data being the session data, the training data according to a question-answer relevance indicated by the label of the session data. (Kumaran)[Abstract] “The processor is configured to generate a machine learning model based on the question, the positive label assigned to the semantically relevant answer,” (Kumaran)[col. 7, lines 41-43]” For each indexed question, the training module 310 generates and trains the machine learning model 312 by providing as input to the machine learning model 312” Kumaran teaches of generating and training a machine learning model for each question, by using question-answer relevance as taught by the claim.
Regarding claim 6:Kumaran teaches:
for each piece of session data, determining the first matching samples as a positive sample, wherein labels of the first matching samples indicate positive question answer relevance (Kumaran)[col.2, lines 47-48]”The subject system assigns a positive label to the semantically relevant answer” Similar to the limitation, Kumaran teaches assigning a positive label to relevant answers to determine semantic relevance.
for each piece of session data, determining the second matching samples as negative samples, wherein labels of the second matching samples indicate negative question answer relevance (Kumaran) [col.2, lines 48-49]” and a negative label to each of the semantically irrelevant answers.” Similar to the limitation, Kumaran teaches assigning a positive label to relevant answers to determine semantic relevance.
and grouping the reference sample, the positive samples and the negative samples into a group of training data. (Kumaran)[col. 7, lines 41-47]For each indexed question, the training module 310 generates and trains the machine learning model 312 by providing as input to the machine learning model 312: the question, the determined semantically relevant answer with its positive label (e.g., 1), and each of the determined semantically irrelevant answers with their respective negative label (e.g., 0). Similar to the limitation, Kumaran teaches of using and grouping the question (reference), positive and negative sample as training data, as taught by the claim. Thus, rendering claim 6 obvious.
Regarding claim 13:
Claim 13 is recited similarly to claim 1. Therefore, it is rejected under the same basis.
Regarding claim 16:
Claim 16 recites similarly to claim 4, therefore it is rejected under the same basis.
Regarding claims 3, 14, 15
Claims 3, 14, and 15 are rejected under 35 U.S.C. 103 as being unpatentable in view of Lixia, in further view of Krishnan and in further view of Kumaran.
Regarding claim 3:
Lixia teaches:
selecting part or all of the categorical data as reference samples (Lixia)[pg.16 pp. 0089]”the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches using labels comprising of string words as references as taught by the claim.
However, Lixia fails to teach:
using each of the reference samples as a target reference sample;
However, Krishnan teaches:
using each of the reference samples as a target reference sample; (Krishnan)[pg.1, pp 0007] “The operations include evaluating a loss function that evaluates a similarity metric between the anchor projected representation and each of the plurality of positive projected representations” Krishnan teaches using an anchor data sample to be compared to, which would then be compared to a positive sample to determine similarity. It would be obvious to a person having ordinary skill of the art that the anchor representation taught by Krishnan would be the target reference sample taught by the claim, thus rendering the claim obvious.
Lixia teaches:
determining the categorical data in a same category (Lixia)[pg.16 pp. 0089]”the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches of determining the category of the data, as taught by the first part of the limitation.
However, Lixia fails to teach:
as the target reference sample as a positive sample associated with the target reference sample;
However, Krishnan teaches:
as the target reference sample as a positive sample associated with the target reference sample; (Krishnan)[pg.1, pp. 0007]” The operations include obtaining an anchor image associated with a first class of a plurality of classes, a plurality of positive images associated with the first class,” Lixia teaches determining categorical data while Krishnan teaches associating a positive sample with the anchor or target reference. When considering the teachings of Lixia with the teachings of Krishnan, it would be obvious to a person having ordinary skill of the art to combine the method of data categorization taught by Lixia, with associating the positive sample with the target sample taught by Krishnan, to determine semantic relevance.
Lixia teaches:
determining the categorical data in a category different from the target reference sample (Lixia)[pg.16 pp. 0089]”the tag data can be tag data used to label the types of sentiment, business, interest, and behavior included in the sample conversation sentences.” Lixia teaches determining categorical data via sorting the data into separate categories.
However, Lixia fails to teach:
as a negative sample associated with the target reference sample;
However, Krishnan teaches:
as a negative sample associated with the target reference sample; (Krishnan)[pg.1, pp .0007]” negative images associated with one or more other classes of the plurality of classes,” Krishnan teaches associating a negative sample with the anchor or target reference.
Krishnan also teaches:
and grouping the target reference sample, the positive sample associated with the target reference sample and the negative sample associated with the target reference sample into a group of training data. (Krishnan)[pg.1, pp. 0007]” wherein the plurality of positive images comprise all images contained within a training batch that are associated with the first class, and wherein the one or more negative anchor images comprise all images contained within the training batch that are not associated with any of the plurality of classes other than the first class.” As taught by the claim, Krishnan teaches of grouping the positive, negative, and plurality of other samples (target or anchor sample data included), to be used as training data, as taught by the claim thus rendering it obvious to a person having ordinary skill of the art.
When considering the teachings of Lixia with the teachings of Krishnan, it would be obvious to a person having ordinary skill of the art to combine the method of data categorization taught by Lixia, with associating the negative sample with the target sample taught by Krishnan, to determine semantic irrelevance. A person having ordinary skill of the art would be motivated to do so that the model can learn new meanings of new knowledge and original knowledge more accurately (Lixia)[Abstract], and to adapt contrastive learning to the fully supervised setting and also enable learning to occur simultaneously across multiple positive examples. (Krishnan)[Abstract].
Regarding claim 14:
Krishnan teaches:
at least one processor; and at least one memory storing a computer program; wherein the computer program is executable by the at least one processor, whereby-and the electronic device is configured to perform the steps of the method according to claim 1 (Krishnan) [pg.1, pp. 0007] “a server computing system[0071]The computing system includes one or more processors and one or more non-transitory computer-readable media that collectively store” Krishnan recites similar components of a computing system, processor, and storage medium, as taught by the claim, thus rendering the claim obvious.
Regarding claim 15:
Krishnan teaches:
A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the steps of the method according to claim 1 (Krishnan)[pg.1, pp. 0007] The computing system includes one or more processors and one or more non-transitory computer-readable media that collectively store:[pg. 6, pp. 0073] The memory 114 can store data 116 and instructions 118 which are executed by the processor 112 to cause the user computing device 102 to perform operations.” Krishnan recites similar components taught by the claim such as a storage medium and a processor to compute the program, thus rendering the claim obvious.
Regarding claims 7 and 17
Claims 7 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Lixia, in view of Kumaran and in further view of McCourt et al., (US10719783B2, referred to as McCourt hereinafter)
Regarding claim 7:
McCourt teaches:
wherein the label of the categorical data is a unitary label, and the label of the session data is a binary label. (McCourt) [col.5, lines 63-65] classification model that can be used to predict a categorical or binary outcome (e.g., true/false or cat/dog) as a function of real-valued inputs. McCourt teaches of using individual identifiers composing of strings as the labels for the categorical data, as well as binary identifiers as taught by the limitation, thus rendering the claim obvious.
Regarding claim 17:
Claim 17 recites similar to claim 7, therefore it is rejected under the same basis.
Regarding claims 8, 10, 18,
Claims 8, 10, 18, are rejected under 35 U.S.C. 103 as being unpatentable in view of Lixia, in further view of Kumaran, in further view of Wei et. Al, (CN112115995A, referred to as Wei hereinafter), and in further view of Cooper et. Al, (US11893774B3, Referred to as Cooper hereinafter).
Regarding claim 8:
Wei teaches:
generating in response to the original data being the data without label, the training data by using data enhancement techniques (Wei)[pg.8 pp.0011]”Extract labeled training images and unlabeled training images from the training dataset.” Wei teaches of using labeled and unlabeled training images, as taught by the claim.
However, Wei fails to teach:
the training data by using data enhancement techniques
However, Cooper teaches:
the training data by using data enhancement techniques (Cooper)[col. 6, lines 26-29]” The image augmentation module 330 generates augmented images by applying an image manipulation function to the labeled training image 400.” Cooper teaches of generating augmented images using image manipulation as taught by the claim.
It would be obvious to a person having ordinary skill of the art to combine the use of data with and without labels as taught by Wei, with the training of data as taught by Cooper. A person having ordinary skill of the art would be motivated to do so to (Wei)[Abstract]” fully utilize effective information of labels, and effectively improve network classification performance.” And (Cooper)[Abstract]predict the training output based on an image training set including the images and the set of augmented images.
Regarding claim 10:
Cooper teaches:
wherein the data without label is a picture, and the data enhancement techniques comprise performing at least one of comprising flipping, mirroring, and cropping on the picture “(Cooper) [col. 1, lines 56-58] developers may augment the image training data with modifications such as flipping, rotating, or cropping the image, which generalize the developed model” Cooper teaches using the same data enhancement techniques that are taught by the claim, thus rendering the claim obvious.
Regarding claim 18:
Claim 18 recites similarly to claim 10, therefore it is rejected under the same basis.
Regarding claim 9
Claim 9 is rejected under 35 U.S.C. 103 as being unpatentable over Lixia, in further view of Wei, in further view of Kumaran, in further view of Cooper and in further view of Krishnan.
Cooper teaches:
wherein the step of generating the training data by using the data enhancement techniques comprises: (Cooper) [col. 6, lines 26-28]” The image augmentation module generates augmented images by applying an image manipulation function to the labeled training image” Cooper teaches of generating augmented images via image manipulation, as taught by the claim.
However, Wei teaches:
using each piece of the data without label as a reference sample; (Wei) [pg.8 pp.0011]” Extract labeled training images and unlabeled training images from the training dataset.” Wei teaches using unlabeled pieces of data as training or reference samples, as taught by the claim.
Cooper teaches:
generating a plurality of… from the reference sample by using the data enhancement techniques;
; (Cooper) [col. 6, lines 26-28]” The image augmentation module generates augmented images by applying an image manipulation function to the labeled training image” Cooper teaches of generating augmented images by applying manipulating or enhancing an image or data, as taught by the claim.
However, Cooper fails to teach:
generating a plurality of positive samples from the reference sample
However, Krishnan teaches:
generating a plurality of positive samples from the reference sample .” (Krishnan) [claim 1] “a plurality of positive images associated with the first class,” Krishnan teaches of generating a plurality of positive images or samples, as taught by the claim, completing the limitation.
It would be obvious to a person having ordinary skill of the art to combine the use of image generation via image augmentation as taught by Cooper with the generation of positive and negative samples as taught by Krishnan. A person having ordinary skill of the art would be motivated to do so to adapt contrastive learning to the fully supervised setting and also enable learning to occur simultaneously across multiple positive examples (Krishnan)[Abstract].
Cooper teaches:
generating a plurality of negative samples from the data without label other than the reference sample by using the data enhancement techniques. (Cooper) [col. 6, lines 26-28] The image augmentation module 330 generates augmented images by applying an image manipulation function to the…training image 400.
However, Cooper fails to teach:
generating a plurality of negative samples
However, Krishnan teaches:
generating a plurality of negative samples (Krishnan) [Claim 1] “one or more negative images associated with one or more other classes of the plurality of classes” Krishnan teaches generating negative images or data samples, as taught by the claim.
It would be obvious to a person having ordinary skill of the art to combine the use of image generation via image augmentation as taught by Cooper with the generation of positive and negative samples as taught by Krishnan. A person having ordinary skill of the art would be motivated to do so to adapt contrastive learning to the fully supervised setting and also enable learning to occur simultaneously across multiple positive examples (Krishnan)[Abstract].
However, Cooper and Krishnan fail to teach:
data without label other than the reference sample
However, Wei teaches:
data without label other than the reference sample by using the data enhancement techniques. (Wei) [pg.8 pp.0011]” Extract labeled training images and unlabeled training images from the training dataset.”
It would be obvious to a person having ordinary skill of the art to combine the use of unlabeled data as taught by Wei, image generation via image augmentation as taught by Cooper with the generation of positive and negative samples as taught by Krishnan. A person having ordinary skill of the art would be motivated to do so to (Wei)[Abstract]” fully utilize effective information of labels, and effectively improve network classification performance.”
Regarding claims 12 and 20:
Claims 12 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Lixia, in further view of Wei, in further view of Kumaran, and in further view of Park et. Al,( US20190354808A1, Park hereinafter)
Regarding claim 12:
Cooper teaches:
wherein the data without label is a sound segment, and the data enhancement techniques comprise performing a random mask operation on the sound segment. (Park) [Claim 19] “obtaining one or more audiographic images that respectively visually represent one or more audio signals;” [pg.3, pp. 0035]” the frequency masking operation can be applied so that f consecutive frequencies [f0; f0+f) are masked” Park teaches altering sound frequencies of an audio graphic image to mask sound segments, as taught by the limitation, therefore rendering the claim obvious.
Regarding claim 20:
Claim 20 recites similarly to claim 12, therefore it is rejected under the same basis.
It would be obvious to a person having ordinary skill of the art to combine the teachings of Lixia, Wei, and Kumaran with the audio masking technique taught by Park. A person having ordinary skill of the art would be motivated to do so to generate augmented training data for machine-learned models. (Park)[pg. 1, pp. 0002]
Regarding claim 11 and 19
Claims 11 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Lixia, in further view of Wei, in further view of Kumaran, and in further view of Reza et. Al, (US12175204B2, referred to as Reza hereinafter).
Regarding claim 11:
Reza teaches:
wherein the data without label is a text, and the data enhancement techniques comprise performing a random mask operation on the text. (Reza)[col. 7, lines 57-58] ”The labels within the prompting templates are then masked with a masking token <Mask>.” [col. 33, lines 40-43]. “The training comprises formulating the task as a masked language modeling problem with the prompting templates” As taught by the claim, Reza recites performing a text masking technique, thus rendering claim 11 obvious.
Regarding claim 19:
Claim 19 recites similar to claim 11, therefore it is rejected under the same basis.
It would be obvious to a person having ordinary skill of the art to combine the use of labeled and unlabeled data as taught by Lixia, with the use of data classification tags as taught by Wei, with the use of positive labels for semantically correct answers as taught by Kumaran with the text masking techniques taught by Reza. A person having ordinary skill of the art would be motivated to do so, by masking the text such that the following prompting templates are generated for the original input (Reza)[col. 7, lines 59-61]
Any inquiry concerning this communication or earlier communications from the examiner should be directed to AYAAN AYAZ SHEIKH whose telephone number is (571)272-4643. The examiner can normally be reached MON-FRI 7:30-5pm.
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, Omar Fernandez can be reached at (571) 272-2589. 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.
/AYAAN AYAZ SHEIKH/Examiner, Art Unit 2128
/OMAR F FERNANDEZ RIVAS/Supervisory Patent Examiner, Art Unit 2128
Prosecution Insights