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 .
Response to Arguments
The present office action is responsive to communication received on 12/17/2025. Claims 1-9, 11-16, 18, and 20 have been amended. Claims 1-20 are currently pending.
Applicant’s amendments file on 12/17/2025 with regards to 35 USC 112, as seen in pages 9-11, have been fully considered but not persuasive.
Applicant’s amendments filed on 12/17/2025 with regards to 35 USC 101, as seen in page 11-12, have been fully considered but not persuasive. With regards to the arguments in view of the change request. Applicant have offered adequate enough support, therefore the rejection with regards to the change request have been withdrawn.
On page 11 of the remarks, the applicant argues:
“…Applicant has amended the pending claims to change "risk- based prediction" to "risk-related prediction." The terms "risk-related prediction" are fully supported and not indefinite because the specification consistently describes outputs that relate to assessing risk associated with implementing a change. It discloses qualitative and numerical risk assessments, binary risky or not risky labels, and numerical scores produced by classifiers and language models, all of which are predictions relating to risk (e.g., "enhanced risk assessment (RA) information" at Paragraph [0037] of the specification; system outputs of predicted RA information at Paragraphs [0086]-[0087] of the specification; RA system outputs including binary and qualitative risk assessments at Paragraphs [0095]-[0096] of the pending specification). "Risk-related prediction" aligns with these disclosures and clarifies that the prediction pertains to risk without altering scope or introducing ambiguity… “
Examiner respectfully disagrees. Examiner asserts the ¶0037, ¶0086-0087, ¶0095-0096, and Figure 6 does not provide sufficient details, such as a algorithmic support, on how the risk assessment machine learning model calculates the risk-related prediction using the first and second inputs. The same argument can be made to an “further-enhanced risk-related prediction”. ¶0039 and ¶0095-0096 does not provide sufficient details, such as algorithmic support, on how risk assessment produces a further-enhanced risk-related prediction using the third and fourth inputs.
On page 11 of the remarks, the applicant argues:
“…Applicant has amended the pending claims to change "a predicted set of documentation textual notes" to "a predicted textual description of one or more outcomes of implementation of the CR." The terms "a predicted textual description of one or more outcomes of implementation of the CR" are likewise supported and definite. The specification defines close notes as the textual description of outcomes of a CR, including notes added during the change deployment process and information from incident reports referencing the change (Paragraphs [0028] of the pending specification). It further teaches training a generative language model on historical change records and close notes so that, in response to a particular CR input, the model predicts outcome information and generates text describing whether the proposed change worked, what portions were successful or not, and progress toward implementation (Paragraph [0086] of the pending specification; see also claim 9 items specifying these outcome elements). The amended phrase directly tracks this disclosed content and therefore is supported and clear…”
Examiner respectfully disagrees. Examiner asserts the ¶0028, ¶0086, and Figure 6 does not provide sufficient details, such as a algorithmic support, on how the generative language learning model to generate the predicted textural description using the deployment activity associated with the IT system or historical change records associated with other deployed IT systems.
Based on the above the arguments, Examiner asserts the independent claims will still be rejected under 112(b).
On page 12 of the remarks, applicant argues that the inclusion of:
“…executing machine learning models that process at least 1000 CRs per month and generate risk predictions using predicted documentation notes produced by a generative language model trained on extensive historical records and deployment-derived close notes… Thes automated ingestion, generation, classification, and retrieval operations over large heterogeneous datasets are beyond what could be mentally performed by a human with reasonable time and resources…”
Examiner respectfully disagrees. Examiner asserts the act of receiving 1000 CRs per month is merely an extra-solution activity of gathering data, even gathering in large quantities, for use in the claimed process, as seen in MPEP § 2106.05(g) & MPEP § 2106.05(h). Insignificant extra-solution activity does not amount to an inventive concept, particularly when the activity is well-understood or conventional. Accordingly, these additional elements do not integrate the abstract idea into a practical application because they do not impose meaningful limits on practicing the abstract ideas. Further the determination of a risk prediction using predicted documentation notes could be done by a human. A human could cross examine historical records and current records to produce a risk prediction. For example, the courts found that wherein using machine learning technology in order to speed up human activity are not eligible under USC 101, Recentive Analytics, Inc. v. Fox Corp. (Fed. Cir. 2025).
Applicant’s arguments and amendments with respect to claims 1, 3, 5, 8, 10-11, 13, 15, 17, and 19-20 stand rejected under 35 U.S.C. § 103 over Huang et al. (US PGPub No. 20250147863-A1) in view of Latha et al. (US PGPub No. 20240013123-A1) and Roytman et al. (US PGPub No. 20240330480-A1) are fully considered and fully persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground of rejection is made in view of Maddila et al. (US PGPub No.20220043779-A1), Ayachitula et al. (US PGPub No. 20240414064-A1), Sawant et al. (US Pat No. 11604626-B1 ), and Sicconi et al. (US PGPub No. 20240112562-A1).
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 non-statutory subject matter claimed invention is directed to an abstract idea without significantly more in light of 2024 Guidance Update on Patent Subject Matter Eligibility, Including on Artificial Intelligence as discussed in 89 FR 58128 at 07/17/2024.
Claim 1 recites a method which appears to be a ‘process’ and one of four statutory categories of invention (Step 1 of the Subject Matter Eligibility Test).
However, the claim as a whole appear to not qualify for a streamlined analysis thus a full eligibility and thus a full eligibility analysis is necessary (Step 2A and Step 2B of the Subject Matter Eligibility Test).
In Step 2A, Prong One, examiners evaluate whether the claim recites a judicial exception, i.e., whether a law of nature, natural phenomenon, or abstract idea is set forth or described in the claim. The claim recites the steps of:
“…executing a risk assessment machine learning model, responsive to receiving a first input and second input…”
“…generate, as an output, risk-related prediction associated with change request (CR)…”
“…executing a generative language machine learning model to generate the second input …”
The steps performing amount to an abstract idea which falls under a judicial exception (Step 2A Prong 1, of Subject Matter Eligibility). Abstract ideas fall in the category. The abstract idea falls in the categories of a mental process, for example evaluation, judgements, and opinion and mathematical concepts (MPEP 2106.04(a)(2) & MPEP 2106.06). For example, the court found that a claim of the application of generic machine learning techniques (e.g., executing risk assessment machine learning model) to the fields of event scheduling and network map creation which merely applies existing technology to a novel database or data environment and insufficient to satisfy step of the Alice inquiry. Additionally, the court found that wherein using machine learning technology in order to speed up human activity (with no improved computer techniques) is not sufficient enough to over 35 USC 101 (page 15) , Recentive Analytics, Inc. v. Fox Corp. (Fed. Cir. 2025).
In Step 2A, Prong Two, examiner determine whether the claim as a whole integrates the judicial exception into a practical application to disqualify abstract as a judicial exception. However, the judicial exception in claim 1 Is not integrated into practical application because the generically recited computer elements do not add meaningful limitation to an abstract idea because they do not add a meaningful limitation to an abstract idea because they amount to simply implementing the abstract idea on a computer. The implementation of using a machine learning techniques in any meaningful to improve the functioning of a computer or another technology without reference to what is well-understood, routine, and conventional activity. The claim do not include additional elements that are sufficient to amount to significantly more than the judicial exception because simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer function that are well-understood, routine and conventional activities previously known to the industry, as discussed in Alice Corp., 573 U.S. at 225, 110 USPQ2d at 1984.
Thus, the analysis concludes is ineligible under 35 U.S.C. § 101 as it is directed to a judicial
exception.
Regarding to claims 3-10:
Claims 3-10 do not add any additional elements than those already disclosed in claim 1, and merely adds further abstract ideas. Furthermore , none of the claims integrate the judicial exception into a practical application.
Claim 11 recites a system which appears to be a ‘machine’ and one of the four statutory categories of invention (Step 1 of the Subject Matter Eligibility Test).
However, the claim as a whole appear to no qualify for a streamlined analysis and thus a full eligibility analysis is necessary (Step 2A and Step 2B of the Subject Matter Eligibility Test).
In Step 2A, Prong One, examiners evaluate whether the claim recites a judicial exception, i.e., whether a law of nature, natural phenomenon, or abstract idea is set forth or described in the claim. The claim recites the steps of:
“…executing a risk assessment machine learning model to, responsive to receiving a first and second input…”
“…generates, as an output, risk-related prediction…”
“…executing a generative language machine learning model to generate the second input …”
The steps performing amount to an abstract idea which falls under a judicial exception (Step 2A Prong 1, of Subject Matter Eligibility). Abstract ideas fall in the category. The abstract idea falls in the categories of a mental process, for example evaluation, judgements, and opinion and mathematical concepts (MPEP 2106.04(a)(2) & MPEP 2106.06). For examples the court found that a claim of the application of generic machine learning techniques to the fields of event scheduling and network map creation which merely applies existing technology to a novel database or data environment and insufficient to satisfy step of the Alice inquiry thus does not create patent eligibility. Additionally, the court found that wherein using machine learning technology in order to speed up human activity (with no improved computer techniques) is not sufficient enough to over 35 USC 101 (page 15) , Recentive Analytics, Inc. v. Fox Corp. (Fed. Cir. 2025).
In Step 2A, Prong Two, examiner determine whether the claim as a whole integrates the judicial exception into a practical application to disqualify abstract as a judicial exception. However, the judicial exception in claim 11 is not integrated into practical application because the generically recited computer elements:
“…a memory electronically coupled to the processor system…”
“…the processor system is operatable to perform processor system operations…”
do not add meaningful limitation to an abstract idea because they do not add a meaningful limitation to an abstract idea because they amount to simply implementing the abstract idea on a computer. The implementation of using a machine learning techniques in any meaningful to improve the functioning of a computer or another technology without reference to what is well-understood, routine, and conventional activity. The claim do not include additional elements that are sufficient to amount to significantly more than the judicial exception because simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer function that are well-understood, routine and conventional activities previously known to the industry, as discussed in Alice Corp., 573 U.S. at 225, 110 USPQ2d at 1984.
Thus, the analysis concludes is ineligible under 35 U.S.C. § 101 as it is directed to a judicial
exception.
Regarding to claims 12-19:
Claims 12-19 do not add any additional elements than those already disclosed in claim 1, and merely adds further abstract ideas. Furthermore, none of the claims integrate the judicial exception into a practical application.
Claim 20 recites computer program product which appears to be a ‘machine’ and one of the four statutory categories of invention (Step 1 of the Subject Matter Eligibility Test).
However, the claim as a whole appear to no qualify for a streamlined analysis and thus a full eligibility analysis is necessary (Step 2A and Step 2B of the Subject Matter Eligibility Test).
In Step 2A, Prong One, examiners evaluate whether the claim recites a judicial exception, i.e., whether a law of nature, natural phenomenon, or abstract idea is set forth or described in the claim. The claim recites the steps of:
“…executing a risk assessment machine learning model to, responsive to receiving a first input and second input…”
“…generate, as an output, a risk-related prediction associated with a change request (CR)…”
“…executing a generative language machine learning model to generate the second input…”
The steps performing amount to an abstract idea which falls under a judicial exception (Step 2A Prong 1, of Subject Matter Eligibility). Abstract ideas fall in the category. The abstract idea falls in the categories of a mental process, for example evaluation, judgements, and opinion and mathematical concepts (MPEP 2106.04(a)(2) & MPEP 2106.06). For examples the court found that a claim of the application of generic machine learning techniques to the fields of event scheduling and network map creation which merely applies existing technology to a novel database or data environment and insufficient to satisfy step of the Alice inquiry thus does not create patent eligibility. Additionally, the court found that wherein using machine learning technology in order to speed up human activity (with no improved computer techniques) is not sufficient enough to over 35 USC 101 (page 15) , Recentive Analytics, Inc. v. Fox Corp. (Fed. Cir. 2025).
In Step 2A, Prong Two, examiner determine whether the claim as a whole integrates the judicial exception into a practical application to disqualify abstract as a judicial exception. However, the judicial exception in claim 20 Is not integrated into practical application because the generically recited computer elements:
“…a computer readable storage medium program instruction operable to instruct a processor system to perform processor system operations…”
do not add meaningful limitation to an abstract idea because they do not add a meaningful limitation to an abstract idea because they amount to simply implementing the abstract idea on a computer. The implementation of using a machine learning techniques in any meaningful to improve the functioning of a computer or another technology without reference to what is well-understood, routine, and conventional activity. The claim do not include additional elements that are sufficient to amount to significantly more than the judicial exception because simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception, e.g., a claim to an abstract idea requiring no more than a generic computer to perform generic computer function that are well-understood, routine and conventional activities previously known to the industry, as discussed in Alice Corp., 573 U.S. at 225, 110 USPQ2d at 1984.
Thus, the analysis concludes is ineligible under 35 U.S.C. § 101 as it is directed to a judicial
exception.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding claims 1, 11, and 20:
“…generates, as an output, risk-related prediction associated with a change request (CR)…”
Indefinite because the claim does not define what a person having ordinary skill in the art would understand under ‘risk-related prediction’ (e.g., score, probability, likelihood). Functional limitations must be supported by structure in the specification, but the application fails to disclose any algorithms, flowcharts, pseudo-code, etc. that offer structural support for this limitation.
“…a predicted textual description of one or more outcomes…”
Indefinite because the claim does not define what constitutes as “predicted textural description” how a predicted set of documentation textual notes is being calculated from the derived deployment activity associated with the IT system or historical change records associated with other deployed IT systems. Functional limitations must be supported by structure in the specification, but the application fails to disclose any algorithm, flowcharts, pseudo-code, etc. that offer structural support for this limitation.
Claims 2-10 and 12-19 do not overcome the rejections of their respective base claims that have been rejected above, and therefore rejected under the same grounds provided to claims 1 and 11.
Regarding claims 4 and 13:
“…determine a further-enhanced risk-related prediction…”
Indefinite because the claim does not define what constitutes as “further-enhanced” nor provides no objective boundaries or metrics by which a person having ordinary skill in the art can perceived what “further-enhanced” may be. It is clear whether this refers score, probability, likelihood. Further there is no algorithmic support within the claims nor specification how of the further-enhanced risk-related prediction is calculated from the third input and fourth input. Functional limitations must be supported by structure in the specification, but the application fails to disclose any algorithms, flowcharts, pseudo-code, etc. that offer structural support for this limitation.
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.
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.
Claims 1, 2, 3, 10, 11, 12, 13, 19, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) and Ayachitula et al. (US PGPub No. 20240414064-A1).
With respect to claim 1, Huang teaches a computer-implemented method comprising: executing a risk assessment machine learning model to, responsive to receiving a first input and a second input, (¶0006: The present invention provides a method of performing code review. The method includes receiving a code change request to merge new code created by developer with an original source code, collecting data associated with each commit in the new code, assessing a risk level of each commit in the new code using an analytical AI and providing code summarization and an initial code review comment of each commit in the new code using generative AI);
generate, as an output, risk-related prediction associated with a change request (CR); (As seen in Figure 3, Table 2, & ¶0030: Last, the hybrid AI solution 150 is configured to send all low-risk commits in the new code NC and the final code review comments of all high-risk commits in the new code NC to the reviewer (step 370)).
wherein the first input comprises the CR received by the risk assessment machine learning model [at a rate of at least 1000 CRs per month;] (As seen in Figure 3, Table 2, & ¶0026-0027: In an embodiment, the developer 10 may create the new code NC by making modifications to the original source code SC, which may consist of bug fix or a new feature. In an embodiment, the developer 10 may submit the new code NC to the code review system 100 via a code change request, such as a pull request (sometimes also referred to as a merge request), in step 210 or 310. In the embodiment depicted in Figure 1, the static scanning tool 120 is configured to statically scan the original source code SC and the new code NC to be evaluated, thereby collecting data DI associated with each commit of the new code NC in step 220 or 320.);
executing a generative language machine learning model to generate the second input, wherein the second input comprises a predicted textual description of one or more outcomes implementation of the CR,(¶0030: As seen in Figure 3, the generative neural network 140 implemented with generative AI is configured is configured to provide the code summarization and the code review comments of the commits in the new code NC in steps 330 and 340. Next, the analytical neural network 130 implemented with the analytic AI is configured to build the predictive model for assessing the risk level of promoting the new code NC based on features of the collected data D1 in steps 350-360.);
Huang does not disclose:
a rate of at least 1000 CRs per month;
However, Maddila teaches a rate of at least 1000 CRs per month; (¶0095: The repositories created and complete at least 1,000 pull requests every month, varied in size, was geographically distributed, and used a variety of programming languages (including Java, C#, Python, Type script, C, C++, SQL, and JS).);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Maddila with regards to a rate of at least 1000 CRs per month to the method of Huang in order to reduce conflicts and duplicative efforts in a large scale environment (Maddila ¶0009-00011).
Huang in view of Maddila does not disclose:
wherein the CR and the risk-related prediction each relates to a requested change for an information technology (IT) system, the IT system comprising one or more deployed systems software, software applications, or hardware that implements the systems software or the software applications; and
the predicted textual description being generated as output from the generative language machine learning model in response to the CR being input into the generative language machine learning model; wherein the predicted textual description is derived from deployment activity associated with the IT system or historical change records associated with other deployed IT systems.
However, Ayachitula wherein the CR and the risk-related prediction each relates to a requested change for an information technology (IT) system, (¶0038: One or more embodiments provide self- learning automated information (IT) change risk prediction. One or more embodiments are configured to assign risk of the change at change creation time in order to predict if the change causes a major incident or an outage. The system is configured to establish relationships between changes via change tickets to an IT environment and incidents via incident tickets in the IT environment);
the IT system comprising one or more deployed systems software, software applications, or hardware that implements the systems software or the software applications; and (¶0147-0161: It is to be understood that although this disclosure includes detailed description on cloud computing implementation of the teachings recited herein are not limited to a cloud computing environment. This cloud model may include at least five characteristics, at least three service models, and at least four deployment models. Infrastructure as a Service (IaaS): the capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications.);
the predicted textual description being generated as output from the generative language machine learning model (¶0144: In one or more embodiments, the machine learning model 220, rule generation algorithm 224, NLP model 228, and/or machine learning model 260 can include various engines/classifiers and/or can be implemented on a neural network. Unsupervised learning can utilize additional learning/training methods including, for example, clustering, anomaly detection, neural networks, deep learning, and the like. ) in response to the CR being input into the generative language machine learning model; (¶0056: The computer system 202 includes a machine learning model 220, which is a client agnostic machine learning model 220, which is a client agnostic machine learning model that has been trained to (only) classify tickets in the IT environment . The machine learning model 220 may be representative of numerous machine learning models 220. The machine learning model 220 classifies the tickets by predicting a label that identifies how to resolve the computer problem associated with ticket. The ticket and its predicted label can be sent to an automated resolution system 222 to automatically resolve the computer problem of the ticket according to the predicted label output from the machine learning model 220.);
wherein the predicted textual description is derived from deployment activity associated with the IT system or historical change records associated with other deployed IT systems. (¶0042: Technical solutions and benefits include system that provides self-learning automated information change risk prediction. In accordance with one or more embodiments, the system can assign a risk of a change at change creation time, which can predict major incidents and outages before they happen, for example, using historical operational data of the IT environment or multiple IT environments as training data. The system can assign risk associated with a change failure using historical operational data as training data. );
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment (Ayachitula ¶0003).
With respect to claim 2, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), but does not disclose wherein the generative language machine learning model has been trained based at least in part on the historical change records associated with the other deployed IT systems. (Ayachitula ¶0042 & ¶0146: In accordance with one or more embodiments, the system can assign a risk of a change at change creation time, which can predict major incidents and outages before they happen, for example, using historical operational data of the IT environment or multiple IT environments as training data. Training datasets (e.g., training data 206, training data 221) can be utilized to train the machine learning algorithms. The training datasets can include historical data of past tickets and the corresponding options/suggestions/resolutions/classification and change categories/verb noun pairs/etc. provided for the respective tickets.);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula of training a model based in part on historical change to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment and to aid to determining the likelihood that the ongoing (or future) change will cause (Ayachitula ¶0042).
With respect to claim 3, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), further comprising presenting, via a computer, at least one of the risk-related prediction and the predicted textual description of one or more outcomes of implementation of the CR. (Ayachitula ¶0093-0094: Figure 7 is example of a graphical user interface. Figure 7 further illustrates dimension predictors for (A) the failure risk. As displayed on the display 119, the typical assessment may determine that the risk is minor as illustrated in the box 712. However, the software application 204 has determined that the overall risk score is critical as illustrated in the box 710 according to one or more embodiments. Further, the software application 204 is configured to generate and display an explanation 720 as the explanation for the overall risk score in accordance with one or more embodiments. Additionally, the software application 204 is configured to generate and display dimension explanations 722 as further explanations of risk individual dimension predictors.);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula of presenting, via computer, at least one of the risk-related prediction and predicted textual description to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment & allow for IT professional (a person in the IT environment) to further review the request (Ayachitula ¶0042 &¶00097 ).
With respect to claim 10, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), wherein the CR comprises at least one member selected from a group consisting of: a textual description of the requested change; a location within the IT system at which the requested change will occur; how the requested change is being applied; one or more entities responsible for applying the requested change; when the requested change is scheduled to be applied; a purpose of the requested change; one or more entities who requested the requested change; and a priority level of the requested change. (Huang ¶0032-0038: The author/developer identification indicates an initial confidence level of the new code NC. Since the author of the original code OC and the developer 10 of the new code NC need to have an in-depth understanding and experience related to the code they are working with, the author/developer identification may include information related whether developer 10 is an authorized person (one or more entities responsible for applying the requested change) to push the new code NC, how many defects the developer 10 has handled previously and the results of his work, whether and how many review comments have been received regarding developer 10 and/or his previous new code submissions, and how-or many quality issues have been raised against the developer 10 or fixed by the developer 10. The author/developer identification may be a preliminary indication on the quality of the new code NC.).
With respect to claim 11, computer system comprising a processor system and a memory electronically coupled to the processor system, wherein the processor system is operable to perform processor system operations comprising; (¶0047: In conclusion, the present invention provides a system and a method capable of enhancing code review efficiency and effectiveness via hybrid AI solution. An analytical AI is implemented for assessing the risk level of each commit in the new code, and a generative AI is implemented for providing code review comments of the commits in the new code. Therefore, the present invention can provide efficient and effective code review with minimal human gatekeeping.);
executing a risk assessment machine learning model to, responsive to receiving a first input and a second input, (¶0006: The present invention provides a method of performing code review. The method includes receiving a code change request to merge new code created by developer with an original source code, collecting data associated with each commit in the new code, assessing a risk level of each commit in the new code using an analytical AI and providing code summarization and an initial code review comment of each commit in the new code using generative AI);
generate, as an output, a risk-related prediction associated with a change request (CR); (As seen in Figure 3, Table 2, & ¶0030: Last, the hybrid AI solution 150 is configured to send all low-risk commits in the new code NC and the final code review comments of all high-risk commits in the new code NC to the reviewer (step 370)).
wherein the first input comprises the CR received by the risk assessment machine learning model at [a rate of at least 1000 CRs per month;] (As seen in Figure 3, Table 2, & ¶0026-0027: In an embodiment, the developer 10 may create the new code NC by making modifications to the original source code SC, which may consist of bug fix or a new feature. In an embodiment, the developer 10 may submit the new code NC to the code review system 100 via a code change request, such as a pull request (sometimes also referred to as a merge request), in step 210 or 310. In the embodiment depicted in Figure 1, the static scanning tool 120 is configured to statically scan the original source code SC and the new code NC to be evaluated, thereby collecting data DI associated with each commit of the new code NC in step 220 or 320.);
and executing a generative language machine learning model to generate the second input, wherein the second input comprises a predicted textual description of one or more outcomes implementation of the CR, (¶0030: As seen in Figure 3, the generative neural network 140 implemented with generative AI is configured is configured to provide the code summarization and the code review comments of the commits in the new code NC in steps 330 and 340. Next, the analytical neural network 130 implemented with the analytic AI is configured to build the predictive model for assessing the risk level of promoting the new code NC based on features of the collected data D1 in steps 350-360.);
Huang does not disclose:
a rate of at least 1000 CRs per month;
However, Maddila teaches a rate of at least 1000 CRs per month; (¶0095: The repositories created and complete at least 1,000 pull requests every month, varied in size, was geographically distributed, and used a variety of programming languages (including Java, C#, Python, Type script, C, C++, SQL, and JS).);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Maddila with regards to a rate of at least 1000 CRs per month to the method of Huang in order to reduce conflicts and duplicative efforts in a large scale environment (Maddila ¶0009-00011).
Huang in view of Maddila does not disclose:
wherein the CR and the risk-related prediction each relates to a requested change for an information technology (IT) system, the IT system comprising one or more deployed systems software, software applications, or hardware that implements the systems software or the software applications;
the predicted textual description being generated as output from the generative language machine learning model in response to the CR being input into the generative language machine learning model; wherein the predicted textual description is derived from deployment activity associated with the IT system or historical change records associated with other deployed IT systems.
However, Ayachitula teaches wherein the CR and the risk-related prediction each relates to a requested change for an information technology (IT) system, (¶0038: One or more embodiments provide self- learning automated information (IT) change risk prediction. One or more embodiments are configured to assign risk of the change at change creation time in order to predict if the change causes a major incident or an outage. The system is configured to establish relationships between changes via change tickets to an IT environment and incidents via incident tickets in the IT environment);
the IT system comprising one or more deployed systems software, software applications, or hardware that implements the systems software or the software applications; (¶0147-0161: It is to be understood that although this disclosure includes detailed description on cloud computing implementation of the teachings recited herein are not limited to a cloud computing environment. This cloud model may include at least five characteristics, at least three service models, and at least four deployment models. Infrastructure as a Service (IaaS): the capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications.);
the predicted textual description being generated as output from the generative language machine learning model (¶0144: In one or more embodiments, the machine learning model 220, rule generation algorithm 224, NLP model 228, and/or machine learning model 260 can include various engines/classifiers and/or can be implemented on a neural network. Unsupervised learning can utilize additional learning/training methods including, for example, clustering, anomaly detection, neural networks, deep learning, and the like. ) in response to the CR being input into the generative language machine learning model; (¶0056: The computer system 202 includes a machine learning model 220, which is a client agnostic machine learning model 220, which is a client agnostic machine learning model that has been trained to (only) classify tickets in the IT environment . The machine learning model 220 may be representative of numerous machine learning models 220. The machine learning model 220 classifies the tickets by predicting a label that identifies how to resolve the computer problem associated with ticket. The ticket and its predicted label can be sent to an automated resolution system 222 to automatically resolve the computer problem of the ticket according to the predicted label output from the machine learning model 220.);
wherein the predicted textual description is derived from deployment activity associated with the IT system or historical change records associated with other deployed IT systems. (¶0042: Technical solutions and benefits include system that provides self-learning automated information change risk prediction. In accordance with one or more embodiments, the system can assign a risk of a change at change creation time, which can predict major incidents and outages before they happen, for example, using historical operational data of the IT environment or multiple IT environments as training data. The system can assign risk associated with a change failure using historical operational data as training data. );
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment (Ayachitula ¶0003).
With respect to claim 12, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above), wherein the generative language machine learning model has been trained based at least in part on the historical change records(Ayachitula ¶0042 & ¶0146: In accordance with one or more embodiments, the system can assign a risk of a change at change creation time, which can predict major incidents and outages before they happen, for example, using historical operational data of the IT environment or multiple IT environments as training data. Training datasets (e.g., training data 206, training data 221) can be utilized to train the machine learning algorithms. The training datasets can include historical data of past tickets and the corresponding options/suggestions/resolutions/classification and change categories/verb noun pairs/etc. provided for the respective tickets.);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula of training a model based in part on historical change to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment and to aid to determining the likelihood that the ongoing (or future) change will cause (Ayachitula ¶0042).
With respect to claim 13, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above), wherein the processor operations further comprise presenting, via a computer, at least one of the risk-related prediction and the predicted textual description of one or more outcomes of implementation of the CR. ( Ayachitula ¶0093-0094: Figure 7 is example of a graphical user interface. Figure 7 further illustrates dimension predictors for (A) the failure risk. As displayed on the display 119, the typical assessment may determine that the risk is minor as illustrated in the box 712. However, the software application 204 has determined that the overall risk score is critical as illustrated in the box 710 according to one or more embodiments. Further, the software application 204 is configured to generate and display an explanation 720 as the explanation for the overall risk score in accordance with one or more embodiments. Additionally, the software application 204 is configured to generate and display dimension explanations 722 as further explanations of risk individual dimension predictors.);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula of presenting the risk-related prediction and predicted textual description to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment & allow for IT professional (a person in the IT environment) to further review the request (Ayachitula ¶0042 &¶00097 ).
With respect to claim 19, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above),wherein the CR comprises at least one member selected from a group consisting of: a textual description of the requested change; a location within the IT system at which the requested change will occur; how the requested change is being applied; one or more entities responsible for applying the requested change; when the requested change is scheduled to be applied; a purpose of the requested change; one or more entities who requested the requested change; and a priority level of the requested change. (Huang ¶0032-0038: The author/developer identification indicates an initial confidence level of the new code NC. Since the author of the original code OC and the developer 10 of the new code NC need to have an in-depth understanding and experience related to the code they are working with, the author/developer identification may include information related whether developer 10 is an authorized person (one or more entities responsible for applying the requested change) to push the new code NC, how many defects the developer 10 has handled previously and the results of his work, whether and how many review comments have been received regarding developer 10 and/or his previous new code submissions, and how-or many quality issues have been raised against the developer 10 or fixed by the developer 10. The author/developer identification may be a preliminary indication on the quality of the new code NC.).
With respect to claim 20, Huang teaches a computer program product comprising a computer readable storage medium storing program instructions operable to instruct a processor system to perform processor system operations comprising: (¶0022-0023: Aspects of the present invention are described herein with reference to the block diagram of the code review system 100 depicted in Figure 1 and the flowcharts depicted in Figures 2-3. It will be understood that each step of the flowcharts depicted in Figures 2-3 may be implemented by computer readable program instructions stored in the computer readable storage medium of the code review system 100. These computer readable program instructions may be provided to a processor of the code review system 100 to produce a machine, such that the instructions, which execute via the processor of the code review system 100, create means for implementing each block in the block diagram depicted in Figure 1 and the functions/acts specified in the flowcharts depicted in Figures 2-3. These computer readable program instructions may also be stored in a computer readable storage medium that can direct the code review system 100 to function in a particular manner, such that the computer readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of each block in the block diagram depicted in Figure 1 and the functions/acts specified in the flowcharts depicted in Figures . 2-3.).
executing a risk assessment machine learning model to, responsive to receiving a first input and a second input, (¶0006: The present invention provides a method of performing code review. The method includes receiving a code change request to merge new code created by developer with an original source code, collecting data associated with each commit in the new code, assessing a risk level of each commit in the new code using an analytical AI and providing code summarization and an initial code review comment of each commit in the new code using generative AI);
generate, as an output, a risk-related prediction associated with a change request (CR); (As seen in Figure 3, Table 2, & ¶0030: Last, the hybrid AI solution 150 is configured to send all low-risk commits in the new code NC and the final code review comments of all high-risk commits in the new code NC to the reviewer (step 370)).
wherein the first input comprises the CR received by the risk assessment machine learning model [at a rate of at least 1000 CRs per month;] (As seen in Figure 3, Table 2, & ¶0026-0027: In an embodiment, the developer 10 may create the new code NC by making modifications to the original source code SC, which may consist of bug fix or a new feature. In an embodiment, the developer 10 may submit the new code NC to the code review system 100 via a code change request, such as a pull request (sometimes also referred to as a merge request), in step 210 or 310. In the embodiment depicted in Figure 1, the static scanning tool 120 is configured to statically scan the original source code SC and the new code NC to be evaluated, thereby collecting data DI associated with each commit of the new code NC in step 220 or 320.);
executing a generative language machine learning model to generate the second input, wherein the second input comprises a predicted textual description of one or more outcomes of implementation of the CR, (¶0030: As seen in Figure 3, the generative neural network 140 implemented with generative AI is configured is configured to provide the code summarization and the code review comments of the commits in the new code NC in steps 330 and 340. Next, the analytical neural network 130 implemented with the analytic AI is configured to build the predictive model for assessing the risk level of promoting the new code NC based on features of the collected data D1 in steps 350-360.);
Huang does not disclose:
a rate of at least 1000 CRs per month;
However, Maddila teaches a rate of at least 1000 CRs per month; (¶0095: The repositories created and complete at least 1,000 pull requests every month, varied in size, was geographically distributed, and used a variety of programming languages (including Java, C#, Python, Type script, C, C++, SQL, and JS).);
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Maddila with regards to a rate of at least 1000 CRs per month to the method of Huang in order to reduce conflicts and duplicative efforts in a large scale environment (Maddila ¶0009-00011).
Huang in view of Maddila does not disclose:
wherein the CR and the risk-related prediction each relates to a requested change for an information technology (IT) system, the IT system comprising one or more deployed systems software, software applications, or hardware that implements the systems software or the software applications; and
the predicted textual description being generated as output from the generative language machine learning model in response to the CR being input into the generative language machine learning model; wherein the predicted textual description is derived from deployment activity associated with the IT system or historical change records associated with other deployed IT systems.
However, Ayachitula wherein the CR and the risk-related prediction each relates to a requested change for an information technology (IT) system, (¶0038: One or more embodiments provide self- learning automated information (IT) change risk prediction. One or more embodiments are configured to assign risk of the change at change creation time in order to predict if the change causes a major incident or an outage. The system is configured to establish relationships between changes via change tickets to an IT environment and incidents via incident tickets in the IT environment);
the IT system comprising one or more deployed systems software, software applications, or hardware that implements the systems software or the software applications; and (¶0147-0161: It is to be understood that although this disclosure includes detailed description on cloud computing implementation of the teachings recited herein are not limited to a cloud computing environment. This cloud model may include at least five characteristics, at least three service models, and at least four deployment models. Infrastructure as a Service (IaaS): the capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications.);
the predicted textual description being generated as output from the generative language machine learning model (¶0144: In one or more embodiments, the machine learning model 220, rule generation algorithm 224, NLP model 228, and/or machine learning model 260 can include various engines/classifiers and/or can be implemented on a neural network. Unsupervised learning can utilize additional learning/training methods including, for example, clustering, anomaly detection, neural networks, deep learning, and the like. ) in response to the CR being input into the generative language machine learning model; (¶0056: The computer system 202 includes a machine learning model 220, which is a client agnostic machine learning model 220, which is a client agnostic machine learning model that has been trained to (only) classify tickets in the IT environment . The machine learning model 220 may be representative of numerous machine learning models 220. The machine learning model 220 classifies the tickets by predicting a label that identifies how to resolve the computer problem associated with ticket. The ticket and its predicted label can be sent to an automated resolution system 222 to automatically resolve the computer problem of the ticket according to the predicted label output from the machine learning model 220.);
wherein the predicted textual description is derived from deployment activity associated with the IT system or historical change records associated with other deployed IT systems. (¶0042: Technical solutions and benefits include system that provides self-learning automated information change risk prediction. In accordance with one or more embodiments, the system can assign a risk of a change at change creation time, which can predict major incidents and outages before they happen, for example, using historical operational data of the IT environment or multiple IT environments as training data. The system can assign risk associated with a change failure using historical operational data as training data. );
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ayachitula to the method of Huang in view of Maddila in order to avoid change induced incidents in the IT environment (Ayachitula ¶0003).
Claims 4 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) , Ayachitula et al. (US PGPub No. 20240414064-A1), and Kwong et al. (US PGPub No. 20210103840-A1 ).
With respect to claim 4, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), but does not disclose responsive to the first input, the second input, a third input, and a fourth input, determine a further-enhanced risk-related prediction associated with the CR; wherein the third input comprises the risk-related prediction; and wherein the fourth input comprises an output from a change record repository.
However, Kwong teaches further comprising: executing a risk assessment system operable to, responsive ( ¶0025-0026: In some implementations of the present solution, the organization uses a prediction tool that is implemented using one or more software processes that execute machine learning algorithm to assess the risk of individual change requests. The parameters of a change request as provided as inputs to the software prediction tool. The ML algorithm processes the input parameters and provides, as an output, an estimate of the probability of success of the change request. For example, the prediction tool can output an estimate showing a 98% success probability for a change request, which indicates that the ML algorithm has determined that there is a 98% probability that the corresponding change request can be implemented without any negative incident. The ML algorithm is trained using historical data about prior change requests, which are stored in a repository, and the algorithm uses this knowledge to determine the output.) to the first input, the second input, a third input, and a fourth input, determine a further-enhanced risk-related prediction associated with the CR; ( In ¶0087 further elaborates wherein information corresponding to one or more prior change requests in the repository are accessed (410). For example, the processor 106 communicates with the repository 120 through the network interface 104, and accesses historical change request dataset, such as information about prior change requests that are stored in the repository 120. The information includes, in one or more entries for each change request, change request identifier 122, input parameters 124, parameter weights 126, predicted values 128 and observed results 130, among other suitable information.).
wherein the third input comprises the risk-related prediction; and (¶0051: In some implementations, a change request entry in the repository 120 includes predicted values 128 which, specifies the success probability that was determined by the change request analysis device 102 using the prediction 114. The predicted values 128 can also include other outputs by the prediction tool 114, such as recommendations for modifications to the change request, outputs by the risk mitigation tool 118, such as alerts for high-risk change request.);
wherein the fourth input comprises an output from a change record repository. (¶0041-0050: The information about the prior change request in repository 120 include input parameters 124. An input parameter can be any value that is needed by the processor 106 to analyze a change request using the prediction tool 114. In some implementations, weights are associated with input parameters to reflect the relative importance of parameters for corresponding change request, and these weights are stored in the repository entry as parameter weights 126. ).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Kwong with regards to applying multiple inputs to the method of Huang in view of Maddila and Ayachitula in order to minimize risk of negative impacts across the organization IT infrastructure due to failed change (Kwong ¶0023).
With respect to claim 14, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above), but does not disclose wherein the processor operations further comprise: executing a risk assessment system operable to, responsive to the first input, the second input, a third input, and a fourth input, determine a further-enhanced risk-related prediction associated with the CR; wherein the third input comprises the risk-related prediction; and wherein the fourth input comprises an output from a change record repository.
However, Kwong teaches wherein the processor operations further comprise: executing a risk assessment system operable to, responsive ( ¶0025-0026: In some implementations of the present solution, the organization uses a prediction tool that is implemented using one or more software processes that execute machine learning algorithm to assess the risk of individual change requests. The parameters of a change request as provided as inputs to the software prediction tool. The ML algorithm processes the input parameters and provides, as an output, an estimate of the probability of success of the change request. For example, the prediction tool can output an estimate showing a 98% success probability for a change request, which indicates that the ML algorithm has determined that there is a 98% probability that the corresponding change request can be implemented without any negative incident. The ML algorithm is trained using historical data about prior change requests, which are stored in a repository, and the algorithm uses this knowledge to determine the output.) to the first input, the second input, a third input, and a fourth input, determine a further-enhanced risk-related prediction associated with the CR; (¶0087 further elaborates wherein information corresponding to one or more prior change requests in the repository are accessed (410). For example, the processor 106 communicates with the repository 120 through the network interface 104, and accesses historical change request dataset, such as information about prior change requests that are stored in the repository 120. The information includes, in one or more entries for each change request, change request identifier 122, input parameters 124, parameter weights 126, predicted values 128 and observed results 130, among other suitable information.).
wherein the third input comprises the risk-related prediction; and (¶0051: In some implementations, a change request entry in the repository 120 includes predicted values 128 which, specifies the success probability that was determined by the change request analysis device 102 using the prediction 114. The predicted values 128 can also include other outputs by the prediction tool 114, such as recommendations for modifications to the change request, outputs by the risk mitigation tool 118, such as alerts for high-risk change request.);
wherein the fourth input comprises an output from a change record repository. (¶0041-0050: The information about the prior change request in repository 120 include input parameters 124. An input parameter can be any value that is needed by the processor 106 to analyze a change request using the prediction tool 114. In some implementations, weights are associated with input parameters to reflect the relative importance of parameters for corresponding change request, and these weights are stored in the repository entry as parameter weights 126. ).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Kwong with regards to applying multiple inputs to the method of Huang in view of Maddila and Ayachitula in order to minimize risk of negative impacts across the organization IT infrastructure due to failed change (Kwong ¶0023).
Claims 5 is rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) , Ayachitula et al. (US PGPub No. 20240414064-A1), and Sicconi et al. (US PGPub No. 20240112562-A1 ).
With respect to claim 5, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), but does not disclose wherein the risk assessment machine learning model comprises the generative language machine learning model.
However, Sicconi teaches wherein the risk assessment machine learning model comprises the generative language machine learning model. (¶0040: Still referring to Figure 1, system 100 may include a unified multimodal neural network. As described in this disclosure, “unified multimodal neural network” refers to the combination of LLM-based digital assistant and risk calculation neural network. In an embodiment, unified multimodal neural network may enhance real-time decision fusion which may be used to seamlessly combine the outputs of LLM-based digital assistant and risk calculation neural network).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Sicconi with regards to risk assessment learning model comprising of a generative language machine learning model to the method of Huang in view of Maddila and Ayachitula in order to further protect the system and better mitigate risks (Sicconi ¶0004).
Claims 6 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) , Ayachitula et al. (US PGPub No. 20240414064-A1), Sicconi et al. (US PGPub No. 20240112562-A1 ), and Lu et al.( US PGPub No.20250291559-A1).
With respect to claim 6, the combination of Huang in view of Maddila, Ayachitula, and Sicconi teaches the method of claim 5 (see rejection of claim 5 above), but does not disclose wherein the generative language machine learning model comprises an encoder.
However, Lu teaches wherein the generative language machine learning model comprises an encoder. (¶0014: As seen in Figure 7B is a block diagram of an example generative language model that includes a transformer-encoder-decoder suitable for use in implementing at least some embodiments of the present disclosure; And as further seen in ¶0081 Various types of LLMs/VLMs/MMLMs/etc. architectures may be implemented in various embodiments. For example, different architectures may be implemented that use different techniques for understanding and generating outputs—such as text, audio, video, image, 2D and/or 3D design or asset data, etc. The LLMs/VLMs/MMLMs/etc. of the present disclosure may include encoder and/or decoder block(s). For example, discriminative or encoder-only models like BERT (Bidirectional Encoder Representations from Transformers) may be implemented for tasks that involve language comprehension such as classification, sentiment analysis, question answering, and named entity recognition.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Lu with regards to the encode to the method of Huang in view of Maddila, Ayachitula, and Sicconi in order to better understand and generate content, such as for translation and summarizations (Lu ¶0081).
With respect to claim 15, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above), but does not disclose wherein: the risk assessment machine learning model comprises the generative language machine learning model; and the generative language machine learning model comprises an encoder.
However, Sicconi teaches wherein: the risk assessment machine learning model comprises the generative language machine learning model; (¶0040: Still referring to Figure 1, system 100 may include a unified multimodal neural network. As described in this disclosure, “unified multimodal neural network” refers to the combination of LLM-based digital assistant and risk calculation neural network. In an embodiment, unified multimodal neural network may enhance real-time decision fusion which may be used to seamlessly combine the outputs of LLM-based digital assistant and risk calculation neural network).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Sicconi with regards to risk assessment learning model comprising of a generative language machine learning model to the method of Huang in view of Maddila and Ayachitula in order to further protect the system and better mitigate risks (Sicconi ¶0004).
Huang in view of Maddila, Ayachitula, and Sicconi does not disclose:
the generative language machine learning model comprises an encoder.
However, Lu teaches the generative language machine learning model comprises an encoder. (¶0014: As seen in Figure 7B is a block diagram of an example generative language model that includes a transformer-encoder-decoder suitable for use in implementing at least some embodiments of the present disclosure; And as further seen in ¶0081 Various types of LLMs/VLMs/MMLMs/etc. architectures may be implemented in various embodiments. For example, different architectures may be implemented that use different techniques for understanding and generating outputs—such as text, audio, video, image, 2D and/or 3D design or asset data, etc. The LLMs/VLMs/MMLMs/etc. of the present disclosure may include encoder and/or decoder block(s). For example, discriminative or encoder-only models like BERT (Bidirectional Encoder Representations from Transformers) may be implemented for tasks that involve language comprehension such as classification, sentiment analysis, question answering, and named entity recognition.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Lu with regards to the encode to the method of Huang in view of Maddila, Ayachitula, and Sicconi in order to better understand and generate content, such as for translation and summarizations (Lu ¶0081).
Claims 7 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) , Ayachitula et al. (US PGPub No. 20240414064-A1) , and Ball et al.( US PGPub No. 20250284579-A1).
With respect to claim 7, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), but does not disclose wherein the risk-related prediction comprises a numerical value that represents a predicted risk of implementing the requested change.
However, Ball teaches wherein the risk-related prediction comprises a numerical value that represents a predicted risk of implementing the requested change. ( ¶0036: Whereas the risk score can be an indication of a confidence that a failure or malfunction will occur, the non-risk score can be the opposite: i.e. a confidence that no failure or malfunction could occur. As an example, the predictive model could predict that a failure will occur by implementing a certain change request. The risk score associated with this outcome prediction could be 95% (i.e. a high likelihood of confidence). Thus, the non-risk score would be 5% in this instance. Put another way, the sum of the risk score and the non-risk score should be 100%, because there are only two possible outcome predictions output by the predictive model.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ball with regards to a numerical value to the method of Huang in view of Maddila and Ayachitula in order to indicate of confidence that a failure or malfunction will occur (Ball ¶0034).
With respect to claim 16, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above),wherein the risk-related prediction comprises a numerical value that represents a predicted risk of implementing the requested change.
However, Ball teaches wherein the risk-related prediction comprises a numerical value that represents a predicted risk of implementing the requested change. ( ¶0036: Whereas the risk score can be an indication of a confidence that a failure or malfunction will occur, the non-risk score can be the opposite: i.e. a confidence that no failure or malfunction could occur. As an example, the predictive model could predict that a failure will occur by implementing a certain change request. The risk score associated with this outcome prediction could be 95% (i.e. a high likelihood of confidence). Thus, the non-risk score would be 5% in this instance. Put another way, the sum of the risk score and the non-risk score should be 100%, because there are only two possible outcome predictions output by the predictive model.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Ball with regards to a numerical value to the method of Huang in view of Maddila and Ayachitula in order to indicate of confidence that a failure or malfunction will occur (Ball ¶0034).
Claims 8 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) , Ayachitula et al. (US PGPub No. 20240414064-A1) , and Sawant et al.( US Pat No. 11604626-B1).
With respect to claim 8, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), but does not disclose wherein the CR comprises at least one of natural language and programming language.
However, Sawant wherein the CR comprises at least one of natural language and programming language. ( Figure 6, is a logical block diagram illustrating coding practice detection, according to some embodiments. As indicated at 620, a practice detection request may be received. The practice detection request 620 may include an identity of a code repository to search, in some embodiments. The request 620 may include the natural language description of the best practice, in some embodiments. As further seen in Figure 7, depicting an implementation of analyzing code, and ¶0072 As indicated at 710, a natural language description of a practice for code written in a programming language may be received, according to some embodiments. The natural language description may be received in a request for detecting coding practices in a code repository.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Sawant with regards to a request comprising at least one natural language and programming language to the method of Huang in view of Maddila and Ayachitula in order to better identify violations and allow for an automated analysis in regards of scaling (Sawant ¶0012-0013).
With respect to claim 17, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above), but does not disclose wherein the CR comprises at least one of natural language and programming language.
However, Sawant wherein the CR comprises at least one of natural language and programming language. ( Figure 6, is a logical block diagram illustrating coding practice detection, according to some embodiments. As indicated at 620, a practice detection request may be received. The practice detection request 620 may include an identity of a code repository to search, in some embodiments. The request 620 may include the natural language description of the best practice, in some embodiments. As further seen in Figure 7, depicting an implementation of analyzing code, and ¶0072 As indicated at 710, a natural language description of a practice for code written in a programming language may be received, according to some embodiments. The natural language description may be received in a request for detecting coding practices in a code repository.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Sawant with regards to a request comprising at least one natural language and programming language to the method of Huang in view of Maddila and Ayachitula in order to better identify violations and allow for an automated analysis in regards of scaling (Sawant ¶0012-0013).
Claims 9 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Huang et al. (US PGPub No. 20250147863-A1) in view of Maddila et al. (US PGPub No. 20220043779-A1) , Ayachitula et al. (US PGPub No. 20240414064-A1) , and Gnaneswaran et al.( US PGPub No. 20210019249-A1).
With respect to claim 9, the combination of Huang in view of Maddila and Ayachitula teaches the method of claim 1 (see rejection of claim 1 above), but does not disclose wherein the predicted textual description of one or more outcome(s) of implementation of the CR includes include at least one member selected from a group consisting of: progress made toward implementing the requested change; whether or not the requested change worked; the portions of the requested change that were successful; and the portions of the requested change that were not successful.
However, Gnaneswaran teaches wherein the predicted textual description of one or more outcome(s) of implementation of the CR includes include at least one member selected from a group consisting of: progress made toward implementing the requested change; whether or not the requested change worked; the portions of the requested change that were successful; and the portions of the requested change that were not successful. (¶0044: For example, the prediction model may determine each, a combination, or at least one of a number of files changed, a number of lines of code changed, a number of patches uploaded for a changeset falls within a range that corresponds to a high risk of failure, a file success rate, an owner success rate, and a review success rate. This determination can be a result of an implementation of a machine learning algorithm that takes the values for feature data, such as the those noted above, as input.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Gnaneswaran with regards to the predicted set of documentation to the method of Huang in view of Maddila and Ayachitula in order to detect potential failures and allow opportunity to rectify root cause of problem stemming from implementation of a changeset (Gnaneswaran ¶0004-0007).
With respect to claim 18, the combination of Huang in view of Maddila and Ayachitula teaches the system of claim 11 (see rejection of claim 11 above), but does not disclose wherein the predicted textual description of one or more outcomes of implementation of the CR include at least one member selected from a group consisting of: progress made toward implementing the requested change; whether or not the requested change worked; the portions of the requested change that were successful; and the portions of the requested change that were not successful.
However, Gnaneswaran teaches wherein the predicted textual description of one or more outcomes of implementation of the CR include at least one member selected from a group consisting of: progress made toward implementing the requested change; whether or not the requested change worked; the portions of the requested change that were successful; and the portions of the requested change that were not successful. (¶0044: For example, the prediction model may determine each, a combination, or at least one of a number of files changed, a number of lines of code changed, a number of patches uploaded for a changeset falls within a range that corresponds to a high risk of failure, a file success rate, an owner success rate, and a review success rate. This determination can be a result of an implementation of a machine learning algorithm that takes the values for feature data, such as the those noted above, as input.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to utilize the teaching of Gnaneswaran with regards to the predicted set of documentation to the method of Huang in view of Maddila and Ayachitula in order to detect potential failures and allow opportunity to rectify root cause of problem stemming from implementation of a changeset (Gnaneswaran ¶0004-0007).
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.
/T.P.V./Examiner, Art Unit 2437
/ALEXANDER LAGOR/Supervisory Patent Examiner, Art Unit 2437