DETAILED ACTION
Claims 23-41 are pending. Claims 23 and 31 have been amended. Claims 1-22 have been previously cancelled
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This final office action is in response to the applicant’s response received on 01/28/2026, for the non-final office action mailed on 10/29/2025.
Examiner’s Notes
Examiner has cited particular columns and line numbers, paragraph numbers, or figures in the references as applied to the claims below for the convenience of the applicant. Although the specified citations are representative of the teachings in the art and are applied to the specific limitations within the individual claim, other passages and figures may apply as well. It is respectfully requested from the applicant, in preparing the responses, to fully consider the references in entirety as potentially teaching all or part of the claimed invention, as well as the context of the passage as taught by the prior art or disclosed by the examiner.
Response to Arguments
Applicant’s arguments filed 01/28/2026 with respect to rejection made under 35 U.S.C. § 103 regarding Lincoln which are not persuasive.
Applicant argues Lincoln does not teach executing a final API flow, wherein the executing results in faster adjustments and risk evaluations over the optimized API flow, the faster adjustments improving the testing of the computer application and the risk evaluations minimizing errors in the testing of the computer application, see applicant’s remarks pp. 16-17. Examiner respectfully disagrees as Lincoln discuss in paragraph [0033] the benefits of understanding API flows and using the understanding to provide opportunities to improve API provisions. Lincoln further teaches in paragraph [0130], “The improvement for the process improves operation of the computer or computer system on which the process is stored and executed. In one embodiment, the improvement of the operation of the computer or computer system on which the process is stored and executed is achieved via one or more of the following features: reducing computer time (i.e., executing the modified process in less time than executing the process prior to being modified) by, for example, reducing the number of API calls; reducing network bandwidth (i.e., using less bandwidth during execution of the modified process than during execution of the process prior to being modified) by, for example, fetching fewer APIs from storage; and simplifying debugging and error analysis of the process by, for example, eliminating the number of interfaces, between APIs of the plurality of APIs, that would otherwise have to be monitored”, when you are reducing the number of API calls you are reducing/minimizing the amount of potential errors when testing a computer application.
Applicant’s arguments filed 01/28/2026 with respect to rejection made under 35 U.S.C. § 103 has been considered but are moot in view of new ground(s) rejection.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 23, 25- 32, 34, 35, 37, 39 and 41 are rejected under 35 U.S.C. 103 as being unpatentable over Subbarayan (US-PGPUB-NO: 2019/0114417 A1), in further view of Arguelles et al. (US-PAT-NO: 10,452,522 B1) hereinafter Arguelles, Anwar et al. (US-PGPUB-NO: 2022/0156633 A1) hereinafter Anwar, Lincoln et al. (US-PGPUB-NO: 2019/0243692 A1) hereinafter Lincoln, Kesarwani (US-PAT-NO: 10,387,295 B1) and Roy et al. (US-PAT-NO: 10,740,164 B1) hereinafter Roy.
As per claim 23, Subbarayan teaches A system for autonomous testing of a computer application, comprising: a non-transitory computer-readable medium configured to store instructions; and at least one processor configured to execute the instructions to perform operations comprising: receiving at least one API flow from a plurality of API flows (see Subbarayan paragraph [0022], showing API calls being received by a destination server) and determining analytic data by inputting the execution data to a clustering model based on a machine leaning technique (see Subbarayan paragraph [0050], showing the use of different machine learning models such as neural network model, random forest model, Bayesian network model, a clustering model and the like and Subbarayan paragraph [0052], showing a machine learning model receiving an API call from the API flow and being able to predict a sequence of API calls (i.e., response data) based on input received).
Subbarayan does not explicitly teach generating test data for executing the at least one API flow; and iteratively determining execution data of the at least one API flow, wherein the determining comprises executing the at least one API flow using the generated test data. However, Arguelles teaches generating test data for executing the at least one API flow; and iteratively determining execution data of the at least one API flow (see Arguelles [column 5, lines 5-20], showing synthetic data being generated to test a sequence of API calls); and wherein the determining comprises executing the at least one API flow using the generated test data (see Arguelles [column 5, lines 54-61], showing the invocation of API calls with test data based on rules to expose web services).
Subbarayan and Arguelles are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies with Arguelles teaches of dynamically generating synthetic data used to test a web service to incorporate invoking an API call to test a web service or an application for better testing coverage or said web service / application.
Subbarayan modified with Arguelles does not explicitly teach iteratively updating a number of API flows in the at least one API flow until a predetermined condition is met; terminating, by a testing module, the iterative updates. However, Anwar teaches iteratively updating a number of API flows in the at least one API flow until a predetermined condition is met; terminating, by a testing module, the iterative updates (see Anwar paragraph [0055], showing iteratively updating only if the change in the size of the model parameter is greater than the difference threshold (i.e., update is being made until difference threshold is not greater signifying a condition being met to stop updating)) (Anwar further teaches in paragraph [0057] the use of a set threshold on parameter change to allow updates which will learn by scaling (i.e., iteratively) and also see Anwar paragraph [0060], showing hitting the max threshold blocks further updates (i.e., terminates)).
Subbarayan, Arguelles and Anwar are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies and Arguelles teaches of dynamically generating synthetic data used to test a web service with Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance to incorporate updating a model based on differences of two datasets and meeting a threshold for said updating and when to stop updating based on the condition.
Subbarayan modified with Arguelles and Anwar do not explicitly teach executing a final API flow, by an API flow generator, wherein the executing results in faster adjustments and risk evaluations over the optimized API flow. However, Lincoln teaches executing a final API flow, by an API flow generator (see Lincoln paragraph [0129], “In one embodiment, improving API provision includes: for an executable process (e.g., method) that uses multiple APIs while being executed, (i) receiving (e.g., from a user) a specification of an improvement to be achieved for the process, (ii) generating a new API that combines a plurality of APIs selected from the multiple APIs”), wherein the executing results in faster adjustments and risk evaluations over the optimized API flow, the faster adjustments improving the testing of the computer application and the risk evaluations minimizing errors in the testing of the computer application (see Lincoln paragraph [0130], “The improvement for the process improves operation of the computer or computer system on which the process is stored and executed. In one embodiment, the improvement of the operation of the computer or computer system on which the process is stored and executed is achieved via one or more of the following features: reducing computer time (i.e., executing the modified process in less time than executing the process prior to being modified) by, for example, reducing the number of API calls; reducing network bandwidth (i.e., using less bandwidth during execution of the modified process than during execution of the process prior to being modified) by, for example, fetching fewer APIs from storage; and simplifying debugging and error analysis of the process by, for example, eliminating the number of interfaces, between APIs of the plurality of APIs, that would otherwise have to be monitored”).
Subbarayan, Arguelles, Anwar and Lincoln are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies, Arguelles teaches of dynamically generating synthetic data used to test a web service and Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance with Lincoln’s teaching of analyzing API usage for improving a computer system to incorporate optimizing API flows in order to improve the reliability of a computer system.
Subbarayan modified with Arguelles, Anwar and Lincoln do not explicitly teach determining, based on an optimized API flow, a regression issue in the test data the regression issue including a cause of a failed API flow and the failed API flow's impact on the testing based on the test data. However, Kesarwani teaches determining, based on an optimized API flow, a regression issue in the test data the regression issue including a cause of a failed API flow and the failed API flow's impact on the testing based on the test data (see Kesarwani [column 10, lines 66-67 and column 11, lines 1-21], “FIG. 4 illustrates an exemplary system and related operations for testing an application using a unit testing framework 430 in a service provider environment 410. For example, the application may be regression tested, stress tested or load tested in order to detect performance problems or determine a robustness level of the application. In one example, various execution orders from multiple clients that are directed to the application may be intermingled. In this case, the application may be tested to verify that although multiple application program interface (API) calls may be performed without an execution guarantee, responses to the API calls are deterministic (i.e., the application produces the same output when provided with the same input) and occur within an expected amount of time. As a non-limiting example, the application that is tested may be related to monitoring customer activity in a computing service environment. For example, the application that is tested may function to collect and track metrics, collect and monitor log files, provide resource utilization, provide application performance, set alarms, etc. with respect to customer activity in the computing service environment. Accordingly, test API calls may be made by executing testing threads 466 to an application or service 490 for monitoring customer activity”).
Subbarayan, Arguelles, Anwar, Lincoln and Kesarwani are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies, Arguelles teaches of dynamically generating synthetic data used to test a web service, Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance and Lincoln’s teaching of analyzing API usage for improving a computer system with Kesarwani’s teaching of application testing using multiple context-aware threads to incorporate regression testing to determine any performance problems or to determine the robustness level of the application based on execution orders of the API calls.
Subbarayan modified with Arguelles, Anwar, Lincoln and Kesarwani do not explicitly teach iteratively updating the number of API flows in the at least one API flow based on the identified regression issue, wherein the iterative updates prevent an erroneous API flow of the at least one API flow from being used. However, Roy teaches iteratively updating the number of API flows in the at least one API flow based on the identified regression issue (see Roy [column 28, lines 31-33], “The learning module 1314 may update the auto-healing corpus 1320 based on an update in the assessment data 204 on a near real-time basis.), wherein the iterative updates prevent an erroneous API flow of the at least one API flow from being used (see Roy [column 28, lines 10-16], “The automatic healing process may further include an auto-healing corpus 1320. The auto-healing corpus 1320 may be the rectification corpus 234 as mentioned above. The auto-healing corpus 1320 may be a collection of different remediation actions, and a repository of steps that may be taken for automated healing of various API's”).
Subbarayan, Arguelles, Anwar, Lincoln, Kesarwani and Roy are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies, Arguelles teaches of dynamically generating synthetic data used to test a web service, Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance, Lincoln’s teaching of analyzing API usage for improving a computer system and Kesarwani’s teaching of application testing using multiple context-aware threads with Roy’s teaching of application programming interface assessment to incorporate regression analysis in order to implement automatic healing to APIs to prevent any errors or any risks that can arise from the use of the APIs.
As per claim 25, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches determining a pattern from the execution data using the clustering model (see Subbarayan paragraph [0049], showing the ML model being used to detect access patterns).
As per claim 26, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches catergorizing the at least one API flow based on the pattern (see Subbarayan paragraph [0049], showing the categorization of malicious activity based on said pattern).
As per claim 27, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches based on the categorizing, determining at least one of: importance levels for the at least on API flow (see Subbarayan paragraph [0083], showing the detected anomalies in transactions can span multiple hierarchical levels of operation of application); impact of a failed API flow in the at least one API flows; or dependency on the at least on API flows.
As per claim 28, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein each API flow of the at least one API flows has a different sequence of composition of the plurality of API flow (see Subbarayan paragraph [0018], showing sequence of API calls which can be sent to different destinations).
As per claim 29, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein, the test data comprises at least one of a tax identification number, a name, an address, a date of birth, or an account balance (see Subbarayan paragraph [0046], showing data input can be account balance).
As per claim 30, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein the execution data is stored in a database (see Subbarayan paragraph [0051], showing the ML model outputting to databases).
As per claim 31, this is the method claim to system claim 23. Therefore, it is rejected for the same reasons as above.
As per claim 32, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches training a prediction model using a supervised machine learning technique (see Subbarayan paragraph [0050], showing the ML model can be a supervised model); determining an inferred response data based on the prediction model (see Subbarayan paragraph [0052], showing a prediction (i.e., inference) being made using the ML model); updating at least one parameter of the prediction model based on the response data (see Subbarayan paragraph [0066], showing the ML model generating an update to the dictionary of symbols / API transactions arising from normal activity).
As per claim 34, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein the prediction model uses initial execution data and initial analytic data as training data (see Subbarayan paragraph [0060-0061], showing the API calls before a first time (i.e., initial execution data) and the indication associated with an API call the second time (i.e., initial analytic data) used to train the machine learning model to identify predicted sequence).
As per claim 35, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein the initial analytic data is determined by inputting execution data into a clustering model (see Subbarayan paragraph [0049], showing the ML model being used to detect access patterns).
As per claim 37, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein the initial execution data is stored in and retrieved from a database (see Subbarayan paragraph [0051], showing the ML model outputting to databases).
As per claim 39, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches wherein the plurality of API flows includes a subset comprising at least one API flow (see Subbarayan paragraph [0018], showing API calls and having one API call (i.e., subset from the API calls) go to one destination and a second API call (i.e., subset from the API calls) going to another destination.
As per claim 41, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy teaches updating the at least one API flow in response to receiving data representing a change in the at least one API flow (see Subbarayan paragraph [0066], showing an update to the dictionary of symbols associated with the patterns of the API transactions).
Claims 24, 33 and 38 are rejected under 35 U.S.C. 103 as being unpatentable over Subbarayan (US-PGPUB-NO: 2019/0114417 A1), Arguelles (US-PAT-NO: 10,452,522 B1), Anwar (US-PGPUB-NO: 2022/0156633 A1), Lincoln (US-PGPUB-NO: 2019/0243692 A1), Kesarwani (US-PAT-NO: 10,387,295 B1) and Roy (US-PAT-NO: 10,740,164 B1), in further view of Trost et al. (US-PGPUB-NO: 2021/0271591 A1) hereinafter Trost.
As per claim 24, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy do not explicitly teach generating test data further comprising generating fake data using a random generator. However, Trost teaches generating test data further comprising generating fake data using a random generator (see Trost paragraph [0037], showing a generator uses random fixed-length vector as input to generate fake data).
Subbarayan, Arguelles, Anwar, Lincoln, Kesarwani, Roy and Trost are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies, Arguelles teaches of dynamically generating synthetic data used to test a web service, Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance, Lincoln’s teaching of analyzing API usage for improving a computer system, Kesarwani’s teaching of application testing using multiple context-aware threads and Roy’s teaching of application programming interface assessment with Trost’s teaching of generating mock data using generative adversarial networks to incorporate generating mock data in order to eliminate the use of production data for testing purposes.
As per claim 33, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani, Roy and Trost teaches wherein the operations further comprise completing training of the prediction model upon determining that the inferred response data meets a predetermined threshold (see Trost paragraph [0014], showing a predetermined limit being used to halt training).
As per claim 38, Subbarayan modified with Arguelles, Anwar, Lincoln, Roy Kesarwani, and Trost teaches wherein the method continued for a predetermined number of iterations (see Trost paragraph [0014], showing a predetermined limit being used for iterating).
Claim 36 is rejected under 35 U.S.C. 103 as being unpatentable over Subbarayan (US-PGPUB-NO: 2019/0114417 A1), Arguelles (US-PAT-NO: 10,452,522 B1), Anwar (US-PGPUB-NO: 2022/0156633 A1), Lincoln (US-PGPUB-NO: 2019/0243692 A1) Kesarwani (US-PAT-NO: 10,387,295 B1) and Roy (US-PAT-NO: 10,740,164 B1), in further view of Dennis et al. (US-PAT-NO: 11,315,092 B1) hereinafter Dennis.
As per claim 36, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy do not explicitly teach wherein the response data comprises at least one message data, error-cause data, or error type data. However, Dennis teaches wherein the response data comprises at least one message data, error-cause data, or error type data (see Dennis [column 11, lines 21-32], showing an API providing a response message that includes data representing an operation code which indicates an outcome of the request message (whether success or failure)).
Subbarayan, Arguelles, Anwar, Lincoln, Kesarwani, Roy and Dennis are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies, Arguelles teaches of dynamically generating synthetic data used to test a web service, Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance, Lincoln’s teaching of analyzing API usage for improving a computer system, Kesarwani’s teaching of application testing using multiple context-aware threads and Roy’s teaching of application programming interface assessment with Dennis’ teaching of an ATM system that uses APIs to interface between two systems to incorporate API outputs responses being related to the messages and whether or not a successful or unsuccessful communication was established.
Claim 40 is rejected under 35 U.S.C. 103 as being unpatentable over Subbarayan (US-PGPUB-NO: 2019/0114417 A1) and Arguelles (US-PAT-NO: 10,452,522 B1), Anwar (US-PGPUB-NO: 2022/0156633 A1), Lincoln (US-PGPUB-NO: 2019/0243692 A1), Kesarwani (US-PAT-NO: 10,387,295 B1) and Roy (US-PAT-NO: 10,740,164 B1), in further view of Corbin, II et al. (US-PGPUB-NO: 2018/0211177 A1) hereinafter Corbin.
As per claim 40, Subbarayan modified with Arguelles, Anwar, Lincoln, Kesarwani and Roy do not explicitly teach wherein the number of API flows within the subset stabilizes after a predetermined number of iterations, terminating the iterations. However, Corbin teaches wherein the number of API flows within the subset stabilizes after a predetermined number of iterations, terminating the iterations (see Corbin paragraph [0241], showing a stop criterion based on a predetermined value for stabilizing iterations).
Subbarayan, Arguelles, Anwar, Lincoln, Kesarwani, Roy and Corbin are analogous art because they are in the same field of endeavor of software development. Therefore, it would have been obvious to one of ordinary skills in the art before the effective filing date of the claimed invention to modify Subbarayan’s teaching of inputting an API call to a machine learning model in order to predict anomalies, Arguelles teaches of dynamically generating synthetic data used to test a web service, Anwar’s teaching of performing federated learning efficiently by optimizing thew size of updates between devices within the system according to model performance, Lincoln’s teaching of analyzing API usage for improving a computer system, Kesarwani’s teaching of application testing using multiple context-aware threads and Roy’s teaching of application programming interface assessment with Corbin’s teaching of generating Bayes net content to incorporate stabilization of data after a predetermined amount of iterations in order to optimize the API traffic being taught in Subbarayan.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Djosic et al. (US-PGPUB-NO: 2021/0152555 A1) teaches detecting unauthorized activity.
Ang (US-PGPUB-NO: 2021/0406711 A1) teaches web application programming interface calls predictions.
Muguda (US-PGPUB-NO: 2014/0282626 A1) teaches application programing interface traffic and caching and enabling equitable bandwidth distribution of the API traffic.
Sriharsha et al. (US-PGPUB-NO: 2022/0036177 A1) teaches data field extraction by a data intake and query system.
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 LENIN PAULINO whose telephone number is (571)270-1734. The examiner can normally be reached Week 1: Mon-Thu 7:30am - 5:00pm Week 2: Mon-Thu 7:30am - 5:00pm and Fri 7:30am - 4:00pm EST.
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, Bradley Teets can be reached on (571) 272-3338. 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.
/LENIN PAULINO/Examiner, Art Unit 2197
/BRADLEY A TEETS/Supervisory Patent Examiner, Art Unit 2197
Prosecution Insights