DETAILED ACTION
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 .
This Office Action has been issued in response to Applicant’s Communication of application S/N 18/435,996 filed on February 05, 2026. Claims 1-20 are pending with the application.
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 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
With respect to claim 1, the limitations directed towards determining, for one or more topics of a plurality of topics, respective stability metrics for individual ones of a plurality of iterations of training of a topic model, wherein the training comprises a first iteration and the plurality of iterations, wherein the plurality of iterations are subsequent to the first iteration, and wherein, for individual topics of the one or more topics, an iteration of the plurality of iterations comprises: selecting a number of elements according to respective probabilities assigned to individual ones of a plurality of elements of a vocabulary of a training set during training of the topic model for the iteration, wherein the selected elements have respective higher probabilities of association with the individual topics than other elements of the plurality of elements that are not selected; and comparing the selected elements of the iteration to previously selected elements of another iteration previous to the iteration, according to a similarity function, to generate the respective stability metrics for the individual topics, is a process that, under its broadest reasonably interpretation, covers performance of these limitations in the mind but for the recitation of generic computer components. That is, nothing in the claim precludes these steps from practically being performed in the mind and/or by a human with pen and paper.
For example, the mention of determining, for one or more topics of a plurality of topics, respective stability metrics for individual ones of a plurality of iterations of training of a topic model, wherein the training comprises a first iteration and the plurality of iterations, wherein the plurality of iterations are subsequent to the first iteration, and wherein, for individual topics of the one or more topics, an iteration of the plurality of iterations comprises: selecting a number of elements according to respective probabilities assigned to individual ones of a plurality of elements of a vocabulary of a training set during training of the topic model for the iteration, wherein the selected elements have respective higher probabilities of association with the individual topics than other elements of the plurality of elements that are not selected; and comparing the selected elements of the iteration to previously selected elements of another iteration previous to the iteration, according to a similarity function, to generate the respective stability metrics for the individual topics, in the context of this claim, encompasses a user mentally determining topic convergence using an iterative algorithm. If a claim limitation, under its broadest reasonable interpretation, covers performance of these limitations in the mind but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
The judicial exception is not integrated into a practical application by additional elements. This claim also does not include additional elements that are sufficient to amount to significantly more than the judicial exception. Claim 1 is not patent eligible.
Claims 8 and 15 recite similar limitations as in claim 1. Therefore claim 8 and 15 are rejected for the same reasons as set forth above. See claim 1 for analysis.
With respects to claims 2, 9, and 16, the limitations are directed towards wherein the similarity function indicates a similarity between the iteration and the other iteration inversely proportional to respective differences between respective probabilities assigned to the selected elements and the previously selected elements. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claims 2, 9, and 16 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
With respects to claims 3, 10, and 17, the limitations are directed towards for individual topics of the one or more topics, comparing respective stability metrics for the iteration and another iteration of the plurality of iterations to generate a convergence metric, wherein the convergence metric indicates a measure of convergence of training of the topic model for the individual topic. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claims 3, 10, and 17 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
With respects to claim 4, the limitations are directed towards wherein an iteration of the plurality of iterations further comprises training the topic model on the training set to assign respective probabilities to individual ones of the plurality of elements of the training set. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claim 4 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
With respects to claims 5, 12, and 19, the limitations are directed towards wherein the number of elements selected is specified to exclude elements with respective assigned probabilities that comprise sampling noise above a threshold value. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claims 5, 12, and 19 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
With respects to claims 6, 11, and 18, the limitations are directed towards wherein, for individual ones of the one or more topics, the respective stability metrics generated for individual ones of the plurality of iterations collective form respective time series stability metrics. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claims 6, 11, and 18 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
With respects to claims 7 and 14, the limitations are directed towards analyzing, for a topic of the one or more topics, the time series stability metrics for the topic to determine that the topic converges for training of the topic model, wherein the analyzing comprises comparing, for a portion of the time series stability metrics, individual values of adjacent ones of the respective time series stability metric to generate a time series convergence metric for the topic, and wherein determining that topic converges for training of the topic model comprises determining that the time series convergence metric meets a threshold value. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claims 7 and 14 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
With respects to claims 13 and 20, the limitations are directed towards herein the number of elements selected is specified as a hyperparameter of the training of the topic model.. These elements further elaborate the abstract idea and the human mind and/or with pen and paper. Therefore, claims 13 and 20 do not recite additional limitations which tie the abstract idea into a practical application and does not amount to significantly more than the identified judicial exception.
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.
Claim(s) 1, 4, 6, 8, 11, 15, and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (U.S. Publication No.: US 20180307680 A1) hereinafter Wu and in view of Kim et al. (U.S. Publication No.: US 20140149417 A1) hereinafter Kim.
As to claim 1:
Wu discloses:
A method, comprising: determining, for one or more topics of a plurality of topics, respective stability metrics for individual ones of a plurality of iterations of training of a topic model [Paragraph 0039 teaches each thread performs iteration processing on the corresponding training text set to calculate a basic LDA model, respectively, and to obtain a probability distribution of basic words to topics and a probability distribution of basic texts to topics of the set.], wherein the training comprises a first iteration and the plurality of iterations, wherein the plurality of iterations are subsequent to the first iteration [Paragraph 0039 teaches each thread performs iteration processing on the corresponding training text set to calculate a basic LDA model, respectively, and to obtain a probability distribution of basic words to topics and a probability distribution of basic texts to topics of the set.], and wherein, for individual topics of the one or more topics, an iteration of the plurality of iterations comprises: selecting a number of elements according to respective probabilities assigned to individual ones of a plurality of elements of a vocabulary of a training set during training of the topic model for the iteration [Paragraph 0039 teaches obtain a probability distribution of basic words to topics and a probability distribution of basic texts to topics of the set… The probability distribution of basic words to topics is a matrix of words to topics with the rows of the matrix being words and the columns of the matrix being topics of implicit calculation.], wherein the selected elements have respective higher probabilities of association with the individual topics than other elements of the plurality of elements that are not selected [Paragraph 0052 teaches complete topic word vectors are obtained from the probability distributions of complete words to topics, i.e. the column vectors in the probability distribution of words to topics. Then, a relevance weight simi(a1, a2) between every two words (a1, a2) in the topic is calculated for each topic word vector i. N clustered topic word vectors can be obtained through training with the basic LDA model and the incremental LDA model, and in each topic word vector, the weight of the word a1 or the word a2 in a topic can be obtained.]; and
Wu discloses all of the limitations as set forth in claim 1 but does not appear to expressly disclose comparing the selected elements of the iteration to previously selected elements of another iteration previous to the iteration, according to a similarity function, to generate the respective stability metric for the individual topic.
Kim discloses:
comparing the selected elements of the iteration to previously selected elements of another iteration previous to the iteration, according to a similarity function, to generate the respective stability metrics for the individual topics [Paragraph 0018 teaches the selected related topics 152 (e.g., Topic 1 and Topic 4, etc.) can be further analyzed to determine a relatedness of each word within the selected related topics 152. For example, for each word a word time series (e.g., word frequency over a period of time) can be determined by determining a quantity of each word that appears in the topic over a particular period of time. Paragraph 0025 teaches it can be advantageous to separate the related words 158 into topic priors 162 based on orientation. Separating the related words 158 into topic priors 162 based on orientation can generate additional topics with a more consistent impact orientation (e.g., consistent positive impact, consistent negative impact, etc.) compared to a previous iteration of generating topics. For example, generating topics using related word priors that have similar orientation can result in topics with a greater relatedness and consistent impact with non-text time series data. Note: Comparing word selections tied to respective different iterations based on a measure of relatedness (similarity) to determine a topic consistency (stability) measure reads on the claims.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu, by incorporating comparing word selections tied to respective different iterations based on a measure of relatedness (similarity) to determine a topic consistency (stability) measure, as taught by Kim (see Paragraph 0025), because both applications are directed to data analysis; incorporating comparing word selections tied to respective different iterations based on a measure of relatedness (similarity) to determine a topic consistency (stability) measure can be advantageous (see Kim Paragraph 0025).
Claims 8 and 15 recite similar limitations as in claim 1. Therefore claim 8 and 15 are rejected for the same reasons as set forth above. See claim 1 for analysis.
As to claim 4:
Wu discloses:
The method of claim 1, wherein an iteration of the plurality of iterations further comprises training the topic model on the training set to assign respective probabilities to individual ones of the plurality of elements of the training set [Paragraph 0039 teaches each thread performs iteration processing on the corresponding training text set to calculate a basic LDA model, respectively, and to obtain a probability distribution of basic words to topics and a probability distribution of basic texts to topics of the set.]
As to claim 6:
Wu and Kim discloses all of the limitations as set forth in claim 1.
Kim also discloses:
The method of claim 1, wherein, for individual ones of the one or more topics, the respective stability metrics generated for individual ones of the plurality of iterations collective form respective time series stability metrics [Paragraph 0018 teaches the selected related topics 152 (e.g., Topic 1 and Topic 4, etc.) can be further analyzed to determine a relatedness of each word within the selected related topics 152. For example, for each word a word time series (e.g., word frequency over a period of time) can be determined by determining a quantity of each word that appears in the topic over a particular period of time. Paragraph 0025 teaches it can be advantageous to separate the related words 158 into topic priors 162 based on orientation. Separating the related words 158 into topic priors 162 based on orientation can generate additional topics with a more consistent impact orientation (e.g., consistent positive impact, consistent negative impact, etc.) compared to a previous iteration of generating topics. For example, generating topics using related word priors that have similar orientation can result in topics with a greater relatedness and consistent impact with non-text time series data. Note: Comparing word selections based on respective different iterations of a time series associated with a measure of relatedness (similarity) to determine a topic consistency (stability) measure reads on the claims.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu, by incorporating comparing word selections based on respective different iterations of a time series associated with a measure of relatedness (similarity) to determine a topic consistency (stability) measure, as taught by Kim (see Paragraph 0025), because both applications are directed to data analysis; incorporating Comparing word selections based on respective different iterations of a time series associated with a measure of relatedness (similarity) to determine a topic consistency (stability) measure can be advantageous (see Kim Paragraph 0025).
Claims 11 and 18 recite similar limitations as in claim 6. Therefore claim 11 and 18 are rejected for the same reasons as set forth above. See claim 6 for analysis.
Claim(s) 2, 9, and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (U.S. Publication No.: US 20180307680 A1) hereinafter Wu and in view of Kim et al. (U.S. Publication No.: US 20140149417 A1) hereinafter Kim, and further in view of Williams (U.S. Patent No.: US 11024274 A1) hereinafter Williams.
As to claim 2:
Wu and Kim discloses all of the limitations as set forth in claim 1 but does not appear to expressly disclose wherein the similarity function indicates a similarity between the iteration and the other iteration inversely proportional to respective differences between respective probabilities assigned to the selected elements and the previously selected elements.
Williams discloses:
The method of claim 1, wherein the similarity function indicates a similarity between the iteration and the other iteration inversely proportional to respective differences between respective probabilities assigned to the selected elements and the previously selected elements [Column 17 Lines 56-60 teaches for a first measure of similarity represented by a value that is inversely proportional to the amount of similarity (such as the minimum of correlation distances), the first criterion may be satisfied when the first measure of similarity is less than the second measure of similarity). Column 18 Lines 7-13 teaches methods may include iterating over multiple different threshold values and combining or otherwise synthesizing the results to provide deeper insight into potential segmentations and ultimately improve the quality of the final segmentation used or accepted. Note: Determining similarity between iterations where the similarity is inversely proportional to other similarity measures associated with segmentations (selected elements) reads on the claims.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu and Kim, by incorporating determining similarity between iterations where the similarity is inversely proportional to other similarity measures associated with segmentations (selected elements), as taught by Williams (see Column 17 Lines 56-60 and Column 18 Lines 7-13), because the three applications are directed to data analysis; incorporating determining similarity between iterations where the similarity is inversely proportional to other similarity measures associated with segmentations (selected elements) provides “measures of similarity” that may be advantageously employed in the present systems (see Williams Column 15 Lines 7-8).
Claims 9 and 16 recite similar limitations as in claim 2. Therefore claim 9 and 16 are rejected for the same reasons as set forth above. See claim 2 for analysis.
Claim(s) 3, 7, 10, 14, and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (U.S. Publication No.: US 20180307680 A1) hereinafter Wu and in view of Kim et al. (U.S. Publication No.: US 20140149417 A1) hereinafter Kim, and further in view of Dunne et al. (U.S. Publication No.: US 20190108270 A1) hereinafter Dunne.
As to claim 3:
Wu and Kim discloses all of the limitations as set forth in claim 1 but does not appear to expressly disclose for individual topics of the one or more topics, comparing respective stability metrics for the iteration and another iteration of the plurality of iterations to generate a convergence metric wherein the convergence metric indicates a measure of convergence of training of the topic model for the individual topic.
Dunne discloses:
The method of claim 1, further comprising, for individual topics of the one or more topics, comparing respective stability metrics for the iteration and another iteration of the plurality of iterations to generate a convergence metric [Paragraph 0091 teaches such iterative repetitions may be appropriate when no patterns of topic convergence and/or topic drift are identified and/or a threshold level is not satisfied. Each iteration may be performed as part of a feedback loop, following the possible identification of topic drift and/or topic convergence and redefining of clusters, in the previous iteration. Paragraph 0092 teaches step 460 involves analysis to identify patterns of changes in MLE scores (or equivalent) of topic keywords over time, which may identify topic drift (corresponding to reducing MLE scores of topic keywords over time) and/or topic convergence (corresponding to increasing MLE scores over time).], wherein the convergence metric indicates a measure of convergence of training of the topic model for the individual topic [Paragraph 0092 teaches step 460 involves analysis to identify patterns of changes in MLE scores (or equivalent) of topic keywords over time, which may identify topic drift (corresponding to reducing MLE scores of topic keywords over time) and/or topic convergence (corresponding to increasing MLE scores over time). Note: MLE score (convergence metric) that changes iteratively in accordance with adjustments to data clusters (data models) reads on the claims.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu and Kim, by incorporating MLE score (convergence metric) that changes iteratively in accordance with adjustments to data clusters (data models), as taught by Dunne (see Paragraph 0091 and 0092), because the three applications are directed to data analysis; incorporating MLE score (convergence metric) that changes iteratively in accordance with adjustments to data clusters (data models) improves data convergence identification (see Dunne Paragraph 0032).
Claims 10 and 17 recite similar limitations as in claim 3. Therefore claim 10 and 17 are rejected for the same reasons as set forth above. See claim 3 for analysis.
As to claim 7:
Wu and Kim discloses all of the limitations as set forth in claim 1 and 6 but does not appear to expressly disclose analyzing, for a topic of the one or more topics, the time series stability metrics for the topic to determine that the topic converges for training of the topic model, wherein the analyzing comprises comparing, for a portion of the time series stability metrics, individual values of adjacent ones of the respective time series stability metric to generate a time series convergence metric for the topic, and wherein determining that topic converges for training of the topic model comprises determining that the time series convergence metric meets a threshold value.
Dune also discloses:
The method of claim 6, further comprising: analyzing, for a topic of the one or more topics, the time series stability metrics for the topic to determine that the topic converges for training of the topic model [Paragraph 0036 teaches comparing the first and second topics to determine a pattern of data convergence of data of the first and second sets may include: determining relevance scores for topics of the primary and secondary data clusters; determining differences between relevance scores of topics of the primary and secondary data clusters; and using the determined differences to identify a pattern of data convergence selected from the group comprising: increasing relevance of one or more topics, indicative of topic convergence; and decreasing relevance of one or more topics, indicative of topic divergence. Paragraph 0088 teaches in order to determine a measure of the increase of the MLE scores of keywords in topic bundles over time (i.e., “topic convergence”). Paragraph 0091 teaches the method may be repeated iteratively using different first (macro) time slices and/or second (micro) time slices. Such iterative repetitions may be appropriate when no patterns of topic convergence and/or topic drift are identified and/or a threshold level is not satisfied. Note: Analyzing a first and second topic at different time (time series) for topic drift (stability) using MLE scores that include (time series stability metrics) to determine convergence reads on the claims.], wherein the analyzing comprises comparing, for a portion of the time series stability metrics, individual values of adjacent ones of the respective time series stability metric to generate a time series convergence metric for the topic [Paragraph 0036 teaches comparing the first and second topics to determine a pattern of data convergence of data of the first and second sets may include: determining relevance scores for topics of the primary and secondary data clusters; determining differences between relevance scores of topics of the primary and secondary data clusters; and using the determined differences to identify a pattern of data convergence selected from the group comprising: increasing relevance of one or more topics, indicative of topic convergence; and decreasing relevance of one or more topics, indicative of topic divergence. Paragraph 0088 teaches in order to determine a measure of the increase of the MLE scores of keywords in topic bundles over time (i.e., “topic convergence”). Paragraph 0091 teaches the method may be repeated iteratively using different first (macro) time slices and/or second (micro) time slices. Such iterative repetitions may be appropriate when no patterns of topic convergence and/or topic drift are identified and/or a threshold level is not satisfied.], and wherein determining that topic converges for training of the topic model comprises determining that the time series convergence metric meets a threshold value [Paragraph 0036 teaches comparing the first and second topics to determine a pattern of data convergence of data of the first and second sets may include: determining relevance scores for topics of the primary and secondary data clusters; determining differences between relevance scores of topics of the primary and secondary data clusters; and using the determined differences to identify a pattern of data convergence selected from the group comprising: increasing relevance of one or more topics, indicative of topic convergence; and decreasing relevance of one or more topics, indicative of topic divergence. Paragraph 0088 teaches in order to determine a measure of the increase of the MLE scores of keywords in topic bundles over time (i.e., “topic convergence”). Paragraph 0091 teaches the method may be repeated iteratively using different first (macro) time slices and/or second (micro) time slices. Such iterative repetitions may be appropriate when no patterns of topic convergence and/or topic drift are identified and/or a threshold level is not satisfied. Note: Analyzing a first and second topic at different time (time series) for topic drift (stability) using MLE scores that include (time series stability metrics) to determine convergence associated relevance scores (metrics), wherein satisfying (meets) a threshold level indicates convergence reads on the claims.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu and Kim, by incorporating analyzing a first and second topic at different time (time series) for topic drift (stability) using MLE scores that include (time series stability metrics) to determine convergence associated relevance scores (metrics), wherein satisfying (meets) a threshold level indicates convergence, as taught by Dunne (see Paragraph 0036, 0088, and 0091), because the three applications are directed to data analysis; incorporating analyzing a first and second topic at different time (time series) for topic drift (stability) using MLE scores that include (time series stability metrics) to determine convergence associated relevance scores (metrics), wherein satisfying (meets) a threshold level indicates convergence improves data convergence identification (see Dunne Paragraph 0032).
Claim 14 recites similar limitations as in claim 7. Therefore claim 14 is rejected for the same reasons as set forth above. See claim 7 for analysis.
Claim(s) 5, 12, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (U.S. Publication No.: US 20180307680 A1) hereinafter Wu and in view of Kim et al. (U.S. Publication No.: US 20140149417 A1) hereinafter Kim, and further in view of Bates et al. (U.S. Publication No.: US 20190286540 A1) hereinafter Bates.
As to claim 5:
Wu and Kim discloses all of the limitations as set forth in claim 1 but does not appear to expressly disclose wherein the number of elements selected is specified to exclude elements with respective assigned probabilities that comprise sampling noise above a threshold value.
Bates discloses:
The method of claim 1, wherein the number of elements selected is specified to exclude elements with respective assigned probabilities that comprise sampling noise above a threshold value [Paragraph 0034 teaches the filters only guarantee to pass on items that are above threshold and items may be sampled to skip one or more filters for various reasons… the probability of sampling may be inversely related to the stable velocity of the below-threshold items (e.g., so the system is not swamped by the class imbalance) and inversely related to the temporal change in the velocity of the above-threshold items (e.g., to take advantage of when there are lulls). Paragraph 0040 teaches the classifier processes and assigns a probability to multiple relevant and non-relevant subclasses. If the probability of relevance is above a predetermined threshold, then it is passed on or communicated to the next module, such as the contextual filter 104. Note: Excluding items based on assigned probabilities above a threshold value to select items that are not passed on reads on the claims.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu and Kim, by incorporating excluding items based on assigned probabilities above a threshold value to select items that are not passed on, as taught by Bates (see Paragraph 0034 and 0040), because the three applications are directed to data analysis; incorporating excluding items based on assigned probabilities above a threshold value to select items that are not passed on provides an improvement on filtering data (see Bates Paragraph 0036).
Claims 12 and 19 recite similar limitations as in claim 5. Therefore claim 12 and 19 are rejected for the same reasons as set forth above. See claim 5 for analysis.
Claim(s) 13 and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wu et al. (U.S. Publication No.: US 20180307680 A1) hereinafter Wu and in view of Kim et al. (U.S. Publication No.: US 20140149417 A1) hereinafter Kim, and further in view of Foroughi et al. (U.S. Publication No.: US 20190303727 A1) hereinafter Foroughi.
As to claim 13:
Wu and Kim discloses all of the limitations as set forth in claim 8 but does not appear to expressly disclose wherein the number of elements selected is specified as a hyperparameter of the training of the topic model.
Foroughi discloses:
The one or more non-transitory, computer-readable storage media of claim 8, wherein the number of elements selected is specified as a hyperparameter of the training of the topic model [Paragraph 0032 teaches the hyperparameter may be the number of latent topics (158) used in the topic modeling.]
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention, to combine the teaching of the cited references and modify the invention as taught by Wu and Kim, by incorporating the hyperparameter may be the number of latent topics, as taught by Foroughi (see Paragraph 0032), because the three applications are directed to data analysis; incorporating the hyperparameter may be the number of latent topics solves the need for need for a trainable, adaptable, and reliable model for information extraction and classification (see Foroughi Paragraph 0001).
Claim 20 recites similar limitations as in claim 13. Therefore claim 20 is rejected for the same reasons as set forth above. See claim 13 for analysis.
Response to Arguments
Applicant presents the following arguments in February 5, 2026 remarks page 13-14:
“…Applicant submits that the claims are directed to statutory subject matter for at least the following reasons.”
Examiner respectfully presents the following response to Applicant’s remarks:
Applicant’s arguments regarding claim 1 USC 101 rejections have been fully considered but they are not persuasive. Regarding independent claim 1, the mention of determining, for one or more topics of a plurality of topics, respective stability metrics for individual ones of a plurality of iterations of training of a topic model, wherein the training comprises a first iteration and the plurality of iterations, wherein the plurality of iterations are subsequent to the first iteration, and wherein, for individual topics of the one or more topics, an iteration of the plurality of iterations comprises: selecting a number of elements according to respective probabilities assigned to individual ones of a plurality of elements of a vocabulary of a training set during training of the topic model for the iteration, wherein the selected elements have respective higher probabilities of association with the individual topics than other elements of the plurality of elements that are not selected; and comparing the selected elements of the iteration to previously selected elements of another iteration previous to the iteration, according to a similarity function, to generate the respective stability metric for the individual topic, in the context of this claim, encompasses a user mentally determining topic convergence using an iterative algorithm. The examiner maintains, if a claim limitation, under its broadest reasonable interpretation, covers performance of these limitations in the mind but for the recitation of generic computer components, then it falls within the “Mental Process” grouping of abstract ideas. Accordingly, the examiner maintains the claim recites an abstract idea.
The judicial exception is not integrated into a practical application by additional elements. This claim also does not include additional elements that are sufficient to amount to significantly more than the judicial exception. Therefore, the examiner maintains claim 1 is not patent eligible.
Applicant presents the following arguments in March 13, 2025 remarks page 8-10:
“…Applicant respectfully submits that the combination of references fails to teach or suggest selecting a number of elements according to respective probabilities, wherein the selected elements have respective higher probabilities of association with the individual topics than other elements of the plurality of elements that are not selected, as claimed.”
Examiner presents the following response to Applicant’s arguments:
Applicant’s arguments have been fully considered but they are not persuasive. Wu’s disclosure sufficiently discloses the current claim 1 recitation of selecting a number of elements according to respective probabilities assigned to individual ones of a plurality of elements of a vocabulary of a training set during training of the topic model for the iteration. Paragraph 0039 teaches obtain a probability distribution of basic words to topics and a probability distribution of basic texts to topics of the set… The probability distribution of basic words to topics is a matrix of words to topics with the rows of the matrix being words and the columns of the matrix being topics of implicit calculation (see Wu Paragraph 0039). The examiner interprets the cited obtaining a probability distribution must include a selection of probability elements making up the probability distribution, wherein the cited probabilities are included in data structure containing words (vocabulary). Therefore, the examiner maintains obtaining a probability distribution of basic words to topics and a probability distribution of basic texts to topics of the set, wherein the probability distribution of basic words to topics is a matrix of words to topics with the rows of the matrix being words and the columns of the matrix being topics of implicit calculation on the claims.
Applicant presents the following arguments in March 13, 2025 remarks page 10-11:
“…Claim 1 features a comparing selected elements of a current iteration to elements of a previous iteration according to a similarity function to generate the respective stability metrics. Applicant respectfully submits that the cited references fail to disclose this feature.”
Examiner presents the following response to Applicant’s arguments:
Applicant’s arguments have been fully considered but they are not persuasive. Wu’s disclosure sufficiently discloses the current claim 1 recitation of selecting a number of elements according to respective probabilities assigned to individual ones of a plurality of elements of a vocabulary of a training set during training of the topic model for the iteration. The selected related topics 152 (e.g., Topic 1 and Topic 4, etc.) can be further analyzed to determine a relatedness of each word within the selected related topics 152. For example, for each word a word time series (e.g., word frequency over a period of time) can be determined by determining a quantity of each word that appears in the topic over a particular period of time (see Paragraph 0018). It can be advantageous to separate the related words 158 into topic priors 162 based on orientation. Separating the related words 158 into topic priors 162 based on orientation can generate additional topics with a more consistent impact orientation (e.g., consistent positive impact, consistent negative impact, etc.) compared to a previous iteration of generating topics. For example, generating topics using related word priors that have similar orientation can result in topics with a greater relatedness and consistent impact with non-text time series data (see Paragraph 0025). The examiner maintains the cited comparing word selections tied to respective different iterations based on a measure of relatedness (similarity) to determine a topic consistency (stability) measure reads on the claims. Further clarification through the amendments to the claim language may aid in differentiating from the current prior art citations.
Conclusion
THIS ACTION IS MADE FINAL. 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 nonprovisional extension fee (37 CFR 1.17(a)) 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 mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to EARL LEVI ELIAS whose telephone number is (571)272-9762. The examiner can normally be reached Monday - Friday (IFP).
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, Sherief Badawi can be reached at 571-272-9782. 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.
/EARL LEVI ELIAS/Examiner, Art Unit 2169 /SHERIEF BADAWI/Supervisory Patent Examiner, Art Unit 2169
Prosecution Insights