Prosecution Insights
Last updated: November 20, 2026
Application No. 19/251,545

Systems and Methods For Detection Of Categorical Draft

Final Rejection §101§103
Filed
Jun 26, 2025
Priority
Dec 03, 2021 — provisional 63/285,997 +1 more
Examiner
MARI VALCARCEL, FERNANDO MARIANO
Art Unit
2159
Tech Center
2100 — Computer Architecture & Software
Assignee
Cisco Technology Inc.
OA Round
2 (Final)
49%
Grant Probability
Moderate
3-4
OA Rounds
2y 1m
Est. Remaining
68%
With Interview

Examiner Intelligence

Grants 49% of resolved cases
49%
Career Allowance Rate
80 granted / 163 resolved
-5.9% vs TC avg
Strong +19% interview lift
Without
With
+18.8%
Interview Lift
resolved cases with interview
Typical timeline
3y 6m
Avg Prosecution
21 currently pending
Career history
205
Total Applications
across all art units

Statute-Specific Performance

§101
15.3%
-24.7% vs TC avg
§103
64.8%
+24.8% vs TC avg
§102
13.6%
-26.4% vs TC avg
§112
6.0%
-34.0% vs TC avg
Black line = Tech Center average estimate • Based on career data from 163 resolved cases

Office Action

§101 §103
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 . Priority Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, 365(c), or 386(c) is acknowledged. The present application is recognized as a continuation of parent Application No. 17/591,535 filed on 12/03/2021. Claim Rejections - 35 USC § 103 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, 13-15 and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over WALTERS et al. (US PGPUB No. 2020/0012900; Pub. Date: Jan. 9, 2020) in view of MOPUR et al. (US PGPUB No. 2021/0365478; Pub. Date: Nov. 25, 2021) and OLGIATI et al. (International Publication No.: WO 2021/067385; Intl. Pub. Date: 08, April, 2021). Regarding independent claim 1, WALTERS discloses a computerized method comprising: parsing received time-series data via a data stream into a plurality of data samples including a first data sample and a second data sample, See FIG. 5A & Paragraphs [0070]-[0072], (Disclosing a system for detecting data drift. FIG. 5 illustrates a method for generating synthetic data for use in another application comprising step 501 of retrieving actual data. Actual data includes streams of information, i.e. parsing received time-series data via a data stream into a plurality of data samples including a first data sample and a second data sample (e.g. a stream of information comprising a plurality of records).) wherein the first data sample includes a first set of input fields corresponding to a first set of data points, and wherein the second data sample includes a second set of input fields corresponding to a second set of data points; See FIG. 5A& Paragraphs [0070]-[0073], (Actual data includes stream of data such as transaction data, human resources data, web log data, web security data, web protocols data, or system logs data. Step 504 of method 500 comprises determining distinct classes of sensitive data portions, i.e. wherein the first data sample includes a first set of input fields corresponding to a first set of data points, and wherein the second data sample includes a second set of input fields corresponding to a second set of data points (e.g. the data stream is parsed into different classes).) and detecting categorical drift has occurred from the first data sample to the second data sample when the degree of divergence satisfies a threshold comparison with a statistical metric of data points of a training data sample. See FIG. 18 & Paragraph [0184], (Determining a difference between predicted data and event data includes determining whether said difference meets or exceeds a threshold difference.) See FIG. 19 & Paragraph [0200], (FIG. 19 illustrates a process for detecting data drift comprising step 1920 of determining a parameter drift threshold indicating a different between an updated model parameter and a baseline model parameter. The difference includes a difference in at least one of a model weight, a model coefficient, a model offset, or another model parameter, i.e. a statistical metric of data points of a training data sample (e.g. the data drift threshold is determined based on characteristics of the model trained on input data).) WALTERS does not disclose the step of generating a first probability distribution based on the first data sample and a second probability distribution based on the second data sample; determining a degree of divergence by providing the first probability distribution and the second probability distribution as input to a distance function; MOPUR discloses the step of generating a first probability distribution based on the first data sample and a second probability distribution based on the second data sample; See Paragraph [0047], (Disclosing a system for updating data models. The method may identify data drift based on changes in densities of a plurality of clusters over a predefined period of time. Densities of the plurality of clusters are prepared by data clustering unit 210 after formation of the plurality of clusters via a histogram algorithm used to determine the distribution of the representative points present within each cluster, i.e. generating a first probability distribution based on the first data sample and a second probability distribution based on the second data sample (e.g. the plurality of clusters determined from input data).) determining a degree of divergence by providing the first probability distribution and the second probability distribution as input to a distance function; See Paragraph [0047], (A cross-correlation may be performed between two related density sets to identify data drift to determine a degree of correlation indicating a similarity between densities across the clusters which would indicate a consistency of the input data over time, i.e. determining a degree of divergence (e.g. by assessing correlations between densities of clusters) by providing the first probability distribution and the second probability distribution as input to a distance function (e.g. the cross-correlation function is applied to the plurality of densities associated with the clusters);) WALTERS and MOPUR are analogous art because they are in the same field of endeavor, correcting model drift. It would have been obvious to anyone having ordinary skill in the art before the effective filing date to modify the system of WALTERS to include the method of clustering an input dataset in order to identify data drift based on changes in densities of said clusters as disclosed by MOPUR. Paragraph [0057] of MOPUR discloses that the system may determine, communicate and use only information corresponding to data drift and outliers in order to save time, bandwidth and computing power when updating data models, which represents an improvement in resource utilization. WALTERS-MOPUR does not disclose the step wherein in response to the detecting of the categorical drift, throttling updating of the training data sample with the second data sample until a number of categorical drift detections for the data stream satisfies a change point threshold, and updating the training data sample with the second data sample in response to the number of categorical drift detections satisfying the change point threshold. OLGIATI discloses the step wherein in response to the detecting of the categorical drift, throttling updating of the training data sample with the second data sample until a number of categorical drift detections for the data stream satisfies a change point threshold, See Paragraph [0025], (Disclosing a system for debugging and profiling of machine learning model training. System comprises a machine learning analysis system for detecting one or more conditions associated with the training of the machine learning model based on an analysis of data associated with the training of said model. Users may configure the system to perform automated problem detection for machine learning models by specifying sampling rates, analysis frequencies and/or thresholds for problem detection in order to detect problems including model drift. See Paragraphs [0026]-[0029], (The analysis process may reduce the quantity and size of tensor data used for analysis, reduction operators are injected into GPUs to produce summarized or aggregated forms of tensor data.) See Paragraph [0101], (The summarized form of tensor data is provided to the analysis system to determine if a particular rule is matched such as by exceeding a threshold, i.e. throttling updating of the training data sample with the second data sample until a number of categorical drift detections for the data stream satisfies a change point threshold (e.g. summarized tensor data is used to train a model as part of a model analysis capable of detecting model drift conditions and initiating actions to improve a model);) and updating the training data sample with the second data sample in response to the number of categorical drift detections satisfying the change point threshold. See Paragraph [0101], (The summarized form of tensor data may be provided to the analysis system which includes rule evaluations representing assertions about the state of the training process. A particular rule may be matched if a particular value in the captured data exceeds a threshold. Note [0057] wherein the thresholds may determine whether a given observation about the model represents a problem that may require intervention, including model drift. The analysis may trigger a retraining of machine learning models based on the problem detection, wherein retraining includes reducing full tensor data to a summarized form as in FIG. 20).) WALTERS-MOPUR and OLGIATI are analogous art because they are in the same field of endeavor, correcting model drift. It would have been obvious to anyone having ordinary skill in the art before the effective filing date to modify the system of WALTERS-MOPUR to include the method of performing a model analysis during a training process as disclosed by OLGIATI. Paragraph [0028] of OLGIATI discloses that the analysis system may provide at least some or all of the following technical improvements: (1) improving the quality of machine learning models in production by automatically detecting common problems and notifying users; (2) improving the quality of machine learning models in production by automatically retraining and redeploying models if problems are detected; (3) improving the performance of inference production with machine learning models by decoupling data collection during inference from data analysis; (4) improving the quality of machine learning models in training by automatically detecting common problems and notifying users; (5) reducing compute resource consumption by discontinuing a model training process after early detection of problems with the training; (6) reducing storage consumption by outputting summarized tensor data from a GPU instead of outputting the full tensor data; (7) reducing network bandwidth consumption by outputting summarized tensor data from a GPU instead of outputting the full tensor data; (8) reducing the latency of operations on tensors by using under utilized GPUs rather than CPUs to perform the operations Regarding dependent claim 4, As discussed above with claim 1, WALTERS-MOPUR-OLGIATI discloses all of the limitations. MOPUR further discloses the step wherein the first set of data points of the first data sample represents are received earlier in time via the data stream than the second set of data points of the second data sample. See FIG. 4 & Paragraph [0043], (FIG. 4 illustrates a plurality of clusters generated via clustering techniques 402, 404, 406 reflecting data points over time. The chart of FIG. 4 illustrates the three clusters in a graph representing point of temperature measured over time wherein clusters 402, 404, 406 are located at distinct time windows, i.e. herein the first set of data points of the first data sample represents are received earlier in time via the data stream (e.g. for example cluster 402) than the second set of data points of the second data sample (e.g. for example cluster 404, 406 which are timestamped at a time past the timestamps of cluster 402).) Regarding dependent claim 6, As discussed above with claim 1, WALTERS-MOPUR-OLGIATI discloses all of the limitations. WALTERS further discloses the step wherein the degree of divergence corresponds to a first input field type, See Paragraph [0191] & [0196], (FIG. 19 comprises step 1908 wherein a baseline data metric of the data profile of the baseline synthetic data is determine wherein the data profile includes a data schema and statistical profile of said synthetic data. Step 1916 comprises detecting data drift based on a difference between a current data metric and a baseline data metric, i.e. wherein the degree of divergence corresponds to a first input field type (e.g. the data metric includes information relating to the schema of the synthetic data, i.e. data fields).) wherein the first data sample and the second data sample include a plurality of input field types. See Paragraph [0144], (Dataset generator 1307 may validate the schema of a synthetic data stream using a JSON validator wherein the schema describes key-value pairs present in a reference data stream. An example schema is described including keys such as "first_name", "last_name" and associated data types, i.e. wherein the first data sample and the second data sample include a plurality of input field types (e.g. the data stream and therefore the distinct classes of data found in said data stream have an associated schema comprising data fields/attributes).) Regarding dependent claim 7, As discussed above with claim 1, WALTERS-MOPUR-OLGIATI discloses all of the limitations. WALTERS further discloses the step wherein each of the training data sample, the first data sample, and the second data sample are received from a same data source. See FIG. 5A & Paragraph [0071], (FIG. 5A depicts process 500 for generating synthetic data from an incoming data stream which is used to train a data model for use in another application by retrieving actual data.) See FIG. 19 & Paragraph [0188], (Process 1900 of FIG. 19 for determining data drift includes receiving model input data which may include actual data, a mix of actual and synthetic data or entirely synthetic data which are derived from an input data stream, i.e. wherein each of the training data sample, the first data sample, and the second data sample are received from a same data source.) Regarding independent claim 8, The claim is analogous to the subject matter of independent claim 1 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding dependent claim 11, The claim is analogous to the subject matter of dependent claim 4 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding dependent claim 13, The claim is analogous to the subject matter of dependent claim 6 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding dependent claim 14, The claim is analogous to the subject matter of dependent claim 7 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding independent claim 15, The claim is analogous to the subject matter of independent claim 1 directed to a device or apparatus and is rejected under similar rationale. Regarding dependent claim 19, The claim is analogous to the subject matter of dependent claim 6 directed to a device or apparatus and is rejected under similar rationale. Regarding dependent claim 20, The claim is analogous to the subject matter of dependent claim 7 directed to a device or apparatus and is rejected under similar rationale. Claim(s) 2-3, 9-10 and 16-17 is/are rejected under 35 U.S.C. 103 as being unpatentable over WALTERS in view of MOPUR and OLGIATTI as applied to claim 1 above, and further in view of JAW (US PGPUB No. 2021/0357702; Pub. Date: Nov. 18, 2021). Regarding dependent claim 2, As discussed above with claim 1, WALTERS-MOPUR-OLGIATI discloses all of the limitations. WALTERS-MOPUR-OLGIATI does not disclose the step of generating the statistical metric of the data points of the training data sample, based on a bootstrap process, for detecting the categorical drift, the statistical metric representing a mean value for the data points of the training data sample. JAW discloses the step of generating the statistical metric of the data points of the training data sample, based on a bootstrap process, for detecting the categorical drift, the statistical metric representing a mean value for the data points of the training data sample. See Paragraphs [0116] & [0118], (Disclosing a system for identifying one or more states of a text string describing an event based on the one or more identified states. The model monitor system is configured to perform data/model integrity checks and detect data drift and accuracy degradation. Data monitored by the model monitor system includes data involved in model training and during production and may comprise statistics that characterize the above datasets including mean, variance and higher order moments of the data sets, i.e. generating the statistical metric of the data points of the training data sample, based on a bootstrap process (e.g. the monitoring process is performed during model training and production), for detecting the categorical drift (e.g. model monitor system may detect data drift), the statistical metric representing a mean value for the data points of the training data sample (e.g. the statistics include mean, variance, etc. of a plurality of datasets).) WALTERS, MOPUR, OLGIATI and JAW are analogous art because they are in the same field of endeavor, correcting model drift. It would have been obvious to anyone having ordinary skill in the art before the effective filing date to modify the system of WALTERS-MOPUR-OLGIATI to include the method of monitoring statistical data for a data model as disclosed by JAW. Paragraph [0054] of JAW discloses that the system may use a plurality of trained classifiers operating in parallel in order to reduce overall prediction latency. The parallel model provides flexibility in re-training, updating or managing an individual predictive model without influencing the performance of other predictive models. Regarding dependent claim 3, As discussed above with claim 2, WALTERS-MOPUR-OLGIATI-JAW discloses all of the limitations. WALTERS further discloses the step wherein the bootstrap process includes determining one or more values associated with an input field type included in the data points of the training data sample, See FIG. 5A & Paragraphs [0070]-[0072], (FIG. 5 illustrates a method for generating synthetic data for use in another application comprising step 501 of retrieving actual data. Actual data includes streams of information.) See Paragraph [0144], (Dataset generator 1307 may validate the schema of a synthetic data stream using a JSON validator wherein the schema descirbe skey-value pairs present in a reference data stream. An example schema is described including keys such as "first_name", "last_name" and associated data types, i.e. wherein the bootstrap process includes determining one or more values associated with an input field type included in the data points of the training data sample (e.g. actual data comprises data having a data schema comprising attributes having types. Actual data is used to train data models).) wherein the data points have the input field type. See Paragraph [0144], (Process 1400 comprises step 1407 of validating attributes of a synthetic data stream generated from actual data by validating the keys present in the synthetic data stream, i.e. wherein the data points have the input field type.) Additionally, JAW further discloses the step of determining the mean value for the data points of the training data sample, See Paragraphs [0117]-[0118], (Data monitored by model monitor system includes statistics that characterize datasets including a mean metric. The system uses the monitored data to perform a process of data/model integrity checks used to detect data drift and accuracy degradation. The process may begin by detecting data drift in training and prediction data by monitoring differences in distributions of training data, text, validation and prediction data, change in distributions of training data, test, validation and prediction data over time, covariates that are causing changes in the prediction output, etc., i.e. determining the mean value for the data points of the training data sample.) Regarding dependent claim 9, The claim is analogous to the subject matter of dependent claim 2 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding dependent claim 10, The claim is analogous to the subject matter of dependent claim 3 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding dependent claim 16, The claim is analogous to the subject matter of dependent claim 2 directed to a device or apparatus and is rejected under similar rationale. Regarding dependent claim 17, The claim is analogous to the subject matter of dependent claim 3 directed to a device or apparatus and is rejected under similar rationale. Claims 5, 12 and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over WALTERS in view of MOPUR and OLGIATI as applied to claim 1 above, and further in view of JAW (US PGPUB No. 2021/0357702; Pub. Date: Nov. 18, 2021). Regarding dependent claim 5, As discussed above with claim 1, WALTERS-MOPUR-OLGIATI discloses all of the limitations. WALTERS-MOPUR-OLGIATI does not disclose the step wherein generating at least one of the first probability distribution and the second probability distribution includes generation by a machine-learning model based on a predetermined probability distribution function. Basel discloses a step wherein generating at least one of the first probability distribution and the second probability distribution includes generation by a machine-learning model based on a predetermined probability distribution function. See Paragraph [0034], (Disclosing a system for performing clustering operations using machine learning techniques. Machine learning classifier 134 is configured to indicate one or more user-defined labels for a plurality of clusters representing real-time time-series data, i.e. wherein generating at least one of the first probability distribution and the second probability distribution includes generation by a machine-learning model based on a predetermined probability distribution function. Machine learning classifier 134 is trained to predict the operating state of a device based on real-time time -series data. Note [0055] wherein the machine learning classifier may identify and compensate for sensor drift via a second clustering operation 220, i.e. wherein generating at least one of the first probability distribution and the second probability distribution includes generation by a machine-learning model (e.g. clusters are generated via machine learning classifier 134) based on a predetermined probability distribution function (e.g. machine learning classifier 134 may perform predictive analysis over the real-time, time-series data).) WALTERS, MOPUR, OLGIATI and Basel are analogous art because they are in the same field of endeavor, correcting data drift. It would have been obvious to anyone having ordinary skill in the art before the effective filing date to modify the system of WALTERS-MOPUR-OLGIATI to include the method of clustering and classifying data using machine learning techniques as disclosed by Basel. Paragraphs [0006] & [0055]-[0056] of Basel disclose that the system may perform subsequent clustering operations to correct for data drift by updating the first machine learning classifier using cluster and label mapping techniques that may reduce the time, expense and labor associated with updating the machine learning classifier. Regarding dependent claim 12, The claim is analogous to the subject matter of dependent claim 5 directed to a non-transitory, computer readable medium and is rejected under similar rationale. Regarding dependent claim 18, The claim is analogous to the subject matter of dependent claim 5 directed to a device or apparatus and is rejected under similar rationale. Response to Arguments Applicant’s arguments with respect to claim(s) 1, 8 and 15 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument. Applicant’s amendments necessitated the new grounds of rejection presented in this Office Action. Applicant’s remarks and amendments with respect to claim(s) 1, 8 and 15 with regard to the rejection of said claims under 35 USC 101 have been considered. The examiner’s view is as follows: Regarding independent claim 1, Claim 1 has been amended to recite the following limitation(s): and in response to the detecting of the categorical drift, throttling updating of the training data sample with the second data sample until a number of categorical drift detections for the data stream satisfies a change point threshold, and updating the training data sample with the second data sample in response to the number of categorical drift detections satisfying the change point threshold. Which recites a step of throttling updating of a training data sample until a number of detections for a data stream satisfy a particular threshold. The process of throttling a data stream represents a step or process that cannot be performed in the human mind neither mentally nor aided by pen & paper. Therefore, claim 1 as currently presented is eligible under 35 USC 101. The corresponding rejection of claim 1 and applicable dependent claims has been withdrawn. Claims 8 and 15 recite identical amendments to claim 1 and are therefore eligible under similar rationale. Therefore, the corresponding rejections have also been withdrawn. Conclusion Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a). A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any 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 Fernando M Mari whose telephone number is (571)272-2498. The examiner can normally be reached Monday-Friday 7am-4pm. 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, Ann J. Lo can be reached at (571) 272-9767. 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. /FMMV/Examiner, Art Unit 2159 /ANN J LO/Supervisory Patent Examiner, Art Unit 2159
Read full office action

Prosecution Timeline

Jun 26, 2025
Application Filed
Jun 15, 2026
Non-Final Rejection mailed — §101, §103
Sep 11, 2026
Applicant Interview (Telephonic)
Sep 11, 2026
Response Filed
Sep 14, 2026
Examiner Interview Summary
Nov 05, 2026
Final Rejection mailed — §101, §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12783548
INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING SYSTEM, INFORMATION PROCESSING METHOD, AND COMPUTER PROGRAM PRODUCT
1y 8m to grant Granted Nov 03, 2026
Patent 12783474
STRUCTURED DATA CONVERSION USING LARGE LANGUAGE MODEL AND FINITE STATE MACHINE
1y 6m to grant Granted Nov 03, 2026
Patent 12743428
METHOD, APPARATUS, AND COMPUTER-READABLE MEDIUM TO EXTRACT A REFERENTIALLY INTACT SUBSET FROM A DATABASE
3y 6m to grant Granted Sep 22, 2026
Patent 12681983
FILE VIEWING METHOD AND FILE VIEWING SYSTEM
2y 7m to grant Granted Jul 14, 2026
Patent 12591588
CATEGORICAL SEARCH USING VISUAL CUES AND HEURISTICS
5y 3m to grant Granted Mar 31, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

Get a prosecution strategy drawn from examiner precedents, rejection analysis, and claim mapping.
Typically takes 5-10 seconds — AI-generated, attorney review required before filing

Prosecution Projections

3-4
Expected OA Rounds
49%
Grant Probability
68%
With Interview (+18.8%)
3y 6m (~2y 1m remaining)
Median Time to Grant
Moderate
PTA Risk
Based on 163 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

Enter your email to receive a magic link. No password needed.

Personal email addresses (Gmail, Yahoo, etc.) are not accepted.

Free tier: 3 strategy analyses per month