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 Amendment
The amendment filed June 8th, 2026, has been entered. Claims 1-2, 3-9, 11-16, and 18-20 remain pending in the application. Claims 3, 10, and 17 have been canceled.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 4-5, 8, 11, 13, 15, and 19 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20230236811 A1 (hereinafter referred to as Sundaresan) and US 20230350657 A1 (hereinafter referred to as Singh), US 11481245 B1 (hereinafter referred to as Oliver) US 20210157577 A1 (hereinafter referred to as Sobran), and Transformers Explained Visually (Part 2): How it works, step-by-step (hereinafter referred to as Doshi).
Regarding claim 1, Sundaresan teaches:
A system for performing multi-platform software code merging and migration using hybrid neural networks, comprising: at least one processing device; at least one memory device; and a module stored in the at least one memory device comprising executable instructions that when executed by the at least one processing device, cause the at least one processing device to ([Abstract], Sundaresan shows an automated system for resolving program merges uses a multi-task neural transformer with attention):
identify one or more source files in a software code development environment, wherein each of the one or more source files comprises software code associated with one or more entity applications that is developed by one or more users in one or more software programming languages (Para. [72], Sundaresan shows “the source code repository may detect a merge conflict and invoke the multi-task neural transformer model to generate a resolution. The programs involved in the merge conflict are obtained from their respective source code repositories (Collectively, block 602). Para. [37], Sundaresan shows a source code repository 202 is a file archive that hosts projects having source code files and other data (e.g., scripts, test cases, etc.) shared by multiple developers. The source code repository 202 may be private and internal to an organization, group, enterprise or user or may be publicly-accessible through the Internet. The source code repository 202 may contain source code files from different domains, such as without limitation, scientific computing, web development, dataflow programming, machine learning, and the like”);
Wherein the transformer based sequence-to-sequence model comprises an encoder and a decoder, wherein the encoder comprises an input embedding layer, a multi-head self-attention layer, a feed forward layer, and the decoder comprises an output embedding layer, a masked multi-head self-attention layer, an encoder-decoder attention layer, a feedforward layer, and a SoftMax layer (Fig. 3A-B; Para. [48], Sundaresan shows “The aggregated decoder output is input to each output layer 235, 237. Edit action output layer 235 includes a linear layer 238 and a SoftMax layer 240.”);
merge the translated code associated with the software code in each of the one or more source files into a unified code, wherein the unified code is in the first software programming language; (Para. [93] Sundaresan “shows detecting a merge conflict between source code snippet A, source code snippet B, and code base O; transforming source code snippet A, source code snippet B, and code base O into a respective abstract syntax tree (AST), wherein source code snippet A and source code snippet B include conflicting modification to code base O; constructing a plurality of aligned AST-node sequences and a plurality of aligned edit sequences, wherein the plurality of aligned AST-node sequences and the plurality of edit sequences are derived from the respective ASTs of source code snippet A, source code snippet B, and code base O; predicting tree editing steps to resolve the merge conflict from a deep learning model, wherein the deep learning model predicts a merge conflict resolution given the plurality of aligned AST-node sequences and the plurality of aligned edit sequences; and applying the predicted tree editing steps to the code base O”);
wherein merging the translated code comprises: determining that the one or more patterns in the translated code causes merge conflicts during deployment (Para. [19-20], Sundaresan shows “The deep learning merge techniques overcome these limitations by using a data-driven approach that learns to predict a merge resolution from real-world merges and merge-conflict data. This data-driven approach considers the semantics of the proposed merger and the syntax of the programming language. Machine learning pertains to the use and development of computer systems that are able to learn and adapt without following explicit instructions, by using algorithms and statistical models to analyze and draw inferences from patterns in data.” Para. [97], Sundaresan shows “The processor is configured to perform acts to: detect a merge conflict between source code component A, source code component B, and code base O, wherein source code component A includes a modification to code base O, wherein source code component B includes a modification to code base O”);
and performing one or more corrective measures to resolve the merge conflicts (Para. [16], Sundaresan shows “automated system for resolving a merge conflict program using a multi-task deep learning model. ”);
Sundaresan does not disclose:
translate the software code in each of the one or more source files that is developed in the one or more software programming languages into a translated code that is in a first software programming language, via a transformer based sequence-to-sequence model;
wherein the transformer based sequence-to-sequence model comprises an encoder and a decoder, wherein the encoder comprises a positional encoding layer, and a residual connection layer
analyzing the translated code to identify one or more patterns associated with one or more known conflicts, via a Siamese neural network;
automatically perform build and packaging of the unified code for deployment using build and packaging tools and requirements associated with the first software programming language, wherein the build and packaging of the unified code does not require multiple build and packaging tools associated with the one or more software programming languages; and automatically deploy the packaged unified code into a software code production environment
and automatically deploy the packaged unified code into a software code production environment.
However, in the analogous art of software translation, Singh teaches:
translate the software code in each of the one or more source files that is developed in the one or more software programming languages into a translated code that is in a first software programming language, via a transformer based sequence-to-sequence model (Para. [27-28]. Singh shows “a code translator 104 may be configured to leverage knowledge of multiple different programming languages in order to aid clients 110-1 to 110-P in translating between programming languages when editing, updating, re-platforming, migrating, or otherwise acting upon their code bases 112-1 to 112-P. For example, code translator 104 may be configured to translate code snippets from one programming language to another, e.g., on the fly or in batches. This may, for instance, enable a developer fluent in a first programming language to view and/or edit source code that was originally written in a second, less-familiar programming language in the first programming language. It may also significantly decrease the time and/or costs associated with migrating code bases between different programming languages. In various implementations, code knowledge system 102 may include a machine learning (“ML” in FIG. 1) database 105 that includes data indicative of one or more trained machine learning models 106-1 to 106-N. These trained machine learning models 106-1 to 106-N may take various forms that will be described in more detail below, including but not limited to BERT (Bidirectional Encoder Representations from Transformers) transformers, GPT (Generative Pre-trained Transformer) transformers, a graph-based network such as a graph neural network (“GNN”), graph attention neural network (“GANN”), or graph convolutional neural network (“GCN”), other types of sequence-to-sequence models and/or encoder-decoders, various flavors of a recurrent neural network (“RNN”, e.g., long short-term memory, or “LSTM”, gate recurrent units, or “GRU”, etc.), and any other type of machine learning model that may be applied to facilitate selected aspects of the present disclosure”);
In addition, in the analogous art of data science and artificial intelligence, Doshi teaches:
wherein the transformer based sequence-to-sequence model comprises an encoder and a decoder, wherein the encoder comprises a positional encoding layer, and a residual connection layer (Section 1: Architecture Overview [First Image]; Doshi shows in the image a transformer architecture common in the art for the purpose of education. Included is a decoder comprising of a positional encoding layer. Section 5: Decoder [Para. 4-6]; Doshi shows “Each of these sub-layers, Self-attention, Encoder-Decoder attention, and Feed-forward, have a residual skip-connection around them, followed by a Layer-Normalization.” )
In addition, in the analogous art of collaborative coding optimization, Sobran teaches:
analyzing the translated code to identify one or more patterns associated with one or more known conflicts, via a Siamese neural network; (Fig. 1, Para. [38], Sobran shows Program 150 generates a defect likelihood of the new commit (step 212). Responsive to program 150 detecting a new commit, program 150 processes a detected commit as described in the steps above. Here, program 150 retrieves associated logs and developmental materials for the detected commit and creates one or more dictionary objects utilizing the aforementioned information. In an embodiment, program 150 removes all unnecessary terms, information, and data, as described in step 204, and vectorizes the resulting commit and associated information. In an embodiment, program 150 creates a dictionary for the new commit. Responsive to program 150 processing and vectorizing the detected commit, program 150 feeds the detected commit into trained model 152. In this embodiment, program 150 determines a log difference, similarity score, and/or defect likelihood for the detected commit utilizing model 152. Bottom of document, [Claim 5-6], wherein the similarity model is a neural network. wherein the neural network is a Siamese network);
In addition, in the analogous art of automated compilations and packaging, Oliver teaches:
automatically perform build and packaging of the unified code for deployment using build and packaging tools and requirements associated with the first software programming language, wherein the build and packaging of the unified code does not require multiple build and packaging tools associated with the one or more software programming languages; and automatically deploy the packaged unified code into a software code production environment. (Para. [28] Oliver shows “a build system 106 comprises a system for automatically determining and executing steps to compile, test, package, and/or deploy a program” Examiner notes the above citation shows a system capable of building and packaging an application. Examiner additionally notes code that has been translated is built and packaged as an application the same way as code that has been developed natively in a software language. The translated code requires the same building and packaging tools as an application that is all built in one language.)
Therefore, it would have been obvious to one of ordinary skill in the art of the claimed invention to incorporate the teachings of Singh into the teachings of Sundaresan to implement “translate the software code in each of the one or more source files that is developed in the one or more software programming languages into a translated code that is in a first software programming language, via a transformer based sequence-to-sequence model“. The modification would have been obvious as one of ordinary skill in the art would be motivated as existing programmers at the entity may lack expertise in the target programming language (Singh, Para. [4]).
In addition, it would have been obvious to one of ordinary skill in the art of the claimed invention to incorporate the teachings of Singh into the teachings of Sundaresan to implement “wherein the transformer based sequence-to-sequence model comprises an encoder and a decoder, wherein the encoder comprises a positional encoding layer, and a residual connection layer“. The modification would have been obvious as one of ordinary skill in the art would be motivated to use a transformer model architecture for the benefits of parallel inputs that give it an advantage over RNN using positional encoding and other layers to retain position information of inputs (Doshi, [Section 2-b: Position Encoding, Para. 2-3]]).
In addition, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Sobran into the teachings of Sundaresan to implement "analyzing the translated code to identify one or more patterns associated with one or more known conflicts, via a Siamese neural network". The modification would have been obvious as one of ordinary skill in the art would be motivated as the traditional approach has significant issues such as requiring humans to label existing logs as problematic, difficult backtracking, and identification of which commit, log, service, or application is causing the issue or problem, and requiring human intervention to determine which commit is defective and proceed with an appropriate remedial action (Sobran, Para. [12]).
In addition, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Oliver into the teachings of Sundaresan as modified to implement “automatically perform build and packaging of the unified code for deployment using build and packaging tools and requirements associated with the first software programming language, wherein the build and packaging of the unified code does not require multiple build and packaging tools associated with the one or more software programming languages; and automatically deploy the packaged unified code into a software code production environment.”. The modification would have been obvious as one of ordinary skill in the art would be motivated as a build tool utilizes special project source code known as a script or a set of scripts for coordinating the build process ... writing and maintaining these scripts becomes a tedious and time-consuming task for developers, as it has to be performed in addition to writing the project source code, thus reducing overall productivity (Oliver, Para. [1]).
Regarding claim 4, Sundaresan as modified teaches claim 1 as cited above and teaches:
wherein the executable instructions cause the at least one processing device to merge the translated code associated with the software code in each of the one or more source files based on one or more historical merge patterns (Para. [36], Sundaresan shows training a multi-task neural transformer with attention model to predict AST-node edit actions and edit labels to resolve a merge conflict. The model is trained from merge tuples and the actions used to resolve the merge conflict (resolving edit actions 232, resolving edit labels 234). A merge tuple (A, B, O) 210 is obtained from a source code repository 202 and includes program A 206 and program B 208, both having changes to a code base O 204. The AST sequence generator 212 parses each source code snippet of the merge tuple 210 into a respective abstract syntax tree, which is traversed in preorder traversal order to generate AST-node sequences which are aligned to generate the aligned AST-node sequences (214, 216, 220, 222) and corresponding edit sequences (218, 224). The aligned AST-node sequences (214, 216, 220, 222) and corresponding edit sequences (218, 224) are applied to the multi-task neural transformer model 226 to learn to predict the edit actions 228 and edit labels 230 that resolve the merge conflict).
Regarding claim 5, Sundaresan as modified teaches claim 1 as cited above however does not disclose:
wherein the executable instructions cause the at least one processing device to: train the transformer based sequence-to-sequence model based on historical snippets of code in each of the one or more software programming languages comprising the first software programming language.
However, in the analogous art of software translation, Singh teaches:
wherein the executable instructions cause the at least one processing device to: train the transformer based sequence-to-sequence model based on historical snippets of code in each of the one or more software programming languages comprising the first software programming language (Para. [29-30], Singh shows in some implementations, code knowledge system 102 may also have access to one or more programming-language-specific corpuses 108-1 to 108-M. In some implementations, these programming-language-specific corpuses 108-1 to 108-M may be used, for instance, to train one or more of the machine learning models 106-1 to 106-N. In some implementations, the programming-language-specific corpuses 108-1 to 108-M may include examples of source code (e.g., entire code bases, libraries, etc.), inline comments, textual metadata associated with source code (e.g., commits), documentation such as textbooks and programming manuals, programming language-specific discussion threads, presentations, academic papers, and so forth. A client 110 that wishes to enable manipulation of its code base 112 in programming language(s) other than that/those used originally to write the source code may establish a relationship with an entity (not depicted in FIG. 1) that hosts code knowledge system 102. When a developer wishes to view/edit a source code snippet of the entity’s code base 112 but is unfamiliar with the native programming language, code translator 104 may provide one or more versions of the source code snippet that is translated to a target programming language preferred by the developer. In some such implementations, code translator 104 may generate the translated source code snippet on the fly, e.g., in real time. In other implementations, code translator 104 may operate, e.g., in a batch mode, to preemptively translate all or selection portions of an entity’s code base 112 into a targeted programming language.).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Singh into the teachings of Sundaresan as modified to implement "train the transformer based sequence-to-sequence model based on historical snippets of code in each of the one or more software programming languages comprising the first software programming language". The modification would have been obvious as one of ordinary skill in the art would be motivated to enable a developer fluent in a first programming language to view and/or edit source code that was originally written in a second, less-familiar programming language in the first programming language (Singh, Para. [27]).
With regards to claim 8, it is a computer program product claim having similar limitations as cited in claim 1 above. Thus, claim 8 is also rejected under the same rationale as cited in the rejection of claim 1 above.
With regards to claim 11, it is a computer program product claim having similar limitations as cited in claim 4 above. Thus, claim 11 is also rejected under the same rationale as cited in the rejection of claim 4 above.
With regards to claim 13, it is a computer program product claim having similar limitations as cited in claim 5 above. Thus, claim 13 is also rejected under the same rationale as cited in the rejection of claim 5 above.
With regards to claim 15, it is a method claim having similar limitations as cited in claim 1 above. Thus, claim 15 is also rejected under the same rationale as cited in the rejection of claim 1 above.
With regards to claim 19, it is a method claim having similar limitations as cited in claim 5 above. Thus, claim 19 is also rejected under the same rationale as cited in the rejection of claim 5 above.
Claim(s) 2, 9, and 16 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20230236811 A1 (hereinafter referred to as Sundaresan) and US 20230350657 A1 (hereinafter referred to as Singh), US 11481245 B1 (hereinafter referred to as Oliver) US 20210157577 A1 (hereinafter referred to as Sobran), and Transformers Explained Visually (Part 2): How it works, step-by-step (hereinafter referred to as Doshi) in further view of US 20230229408 A1 (hereinafter referred to as Gupta).
Regarding claim 2, Sundaresan as modified teaches claim 1 as cited above, but does not disclose:
wherein merging the translated code associated with the software code in each of the one or more source files comprises: analyzing the translated code to detect duplicate code snippets in the translated code;
and in response to detecting the duplicate code snippets in the translated code, isolating the duplicate code snippets from the translated code.
However, in the analogous art of software code optimization, Gupta teaches:
wherein merging the translated code associated with the software code in each of the one or more source files comprises: analyzing the translated code to detect duplicate code snippets in the translated code (Para. [43] Gupta shows “the developer may receive reports from the code evaluation system 400 as any new code written by that developer is committed” Para. [47] Gupta shows “parsing code associated with the code request; storing, in a staging database, a portion of the code; traversing a codebase to identify any code in the codebase that matches the portion of the code”;);
and in response to detecting the duplicate code snippets in the translated code, isolating the duplicate code snippets from the translated code (Para. [43] Gupta shows the developer may receive reports from the code evaluation system 400 as any new code written by that developer is committed to the master DB. Para. [47] Gupta shows parsing code associated with the code request; storing, in a staging database, a portion of the code; traversing a codebase to identify any code in the codebase that matches the portion of the code; and when code is found in the codebase that matches the portion of the code, incrementing a green count and pushing the portion of the code to a redundant code bin, and when no code is found in the codebase that matches the portion of the code, incrementing a red counter and updating the codebase to include the portion of the code).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Gupta into the teachings of Sundaresan as modified to implement “wherein merging the translated code associated with the software code in each of the one or more source files comprises: analyzing the translated code to detect duplicate code snippets in the translated code; and in response to detecting the duplicate code snippets in the translated code, isolating the duplicate code snippets from the translated code”. The modification would have been obvious as one of ordinary skill in the art would be motivated to enable the identification of redundant, or possibly redundant, code and thereby improve the speed and efficiency of development processes by reducing, or eliminating, development of redundant code (Gupta, Para. [11]).
With regards to claim 9, it is a computer program product claim having similar limitations as cited in claim 2 above. Thus, claim 9 is also rejected under the same rationale as cited in the rejection of claim 2 above.
With regards to claim 16, it is a method claim having similar limitations as cited in claim 2 above. Thus, claim 16 is also rejected under the same rationale as cited in the rejection of claim 2 above.
Claim(s) 6, 14, and 20 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20230236811 A1 (hereinafter referred to as Sundaresan) and US 20230350657 A1 (hereinafter referred to as Singh), US 11481245 B1 (hereinafter referred to as Oliver) US 20210157577 A1 (hereinafter referred to as Sobran), and Transformers Explained Visually (Part 2): How it works, step-by-step (hereinafter referred to as Doshi) in further view of US 20230196230 A1 (hereinafter referred to as Cella).
Regarding claim 6, Sundaresan as modified teaches claim 1 as cited above, but does not disclose:
wherein the executable instructions cause the at least one processing device to: provide a virtual environment to the one or more users to allow the one or more users to perform one or more software code related operations associated with the one or more entity applications.
However, in the analogous art of digital twin data visualization, Cella teaches:
wherein the executable instructions cause the at least one processing device to: provide a virtual environment to the one or more users to allow the one or more users to perform one or more software code related operations associated with the one or more entity applications (Para. [4413], Cella shows “a CIO digital twin 60632 may assist in virtual monitoring and testing in a virtual environment to test implementations, changes, reconfigurations, the introduction and/or removal of components and other assets, and the like”).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Cella into the teachings of Sundaresan as modified to implement “wherein the executable instructions cause the at least one processing device to: provide a virtual environment to the one or more users to allow the one or more users to perform one or more software code related operations associated with the one or more entity applications”. The modification would have been obvious as one of ordinary skill in the art would be motivated to allow a user to identify events and consequences that occur as a result of multiple simultaneous or related faults or other events (Cella, Para. [4317]).
With regards to claim 14, it is a computer program product claim having similar limitations as cited in claim 6 above. Thus, claim 14 is also rejected under the same rationale as cited in the rejection of claim 6 above.
With regards to claim 20, it is a method claim having similar limitations as cited in claim 6 above. Thus, claim 20 is also rejected under the same rationale as cited in the rejection of claim 6 above.
Claim(s) 7, 12, and 18 is/are rejected under 35 U.S.C. 103 as being unpatentable over US 20230236811 A1 (hereinafter referred to as Sundaresan) and US 20230350657 A1 (hereinafter referred to as Singh), US 11481245 B1 (hereinafter referred to as Oliver) US 20210157577 A1 (hereinafter referred to as Sobran), and Transformers Explained Visually (Part 2): How it works, step-by-step (hereinafter referred to as Doshi) in further view of US 20230196230 A1 (hereinafter referred to as Cella) and US 20250117201 A1 (hereinafter referred to as Goloubew).
Regarding claim 7, Sundaresan as modified teaches claim 6 as cited above and does not disclose:
wherein the executable instructions cause the at least one processing device to: provide a digital twin of software code environments comprising the software code development environment, software code testing environment,
and the software code production environment;
allow the one or more users to perform one or more software code related operations associated with the one or more entity applications in the digital twin; determine that the one or more software code related operations are successful; and implement the one or more software code related operations performed in the digital twin in the software code environments.
However, in the analogous art of digital twin data visualization, Cella teaches:
wherein the executable instructions cause the at least one processing device to: provide a digital twin of software code environments comprising the software code development environment, software code testing environment (Para. [4413], Cella shows a CIO digital twin 60632 may assist in virtual monitoring and testing in a virtual environment. Para. [4398] Cella shows a CTO digital twin 60630 may be configured to allow a user to research, create, track and report on a technology, development),
allow the one or more users to perform one or more software code related operations associated with the one or more entity applications in the digital twin (Para. [4398] a CTO digital twin 60630 may be configured to allow a user to research, create, track and report on a technology, development, and/or technology or engineering department initiative including, but not limited to, a new product development, update, enhancement, replacement, upgrade, or the like);
determine that the one or more software code related operations are successful (Para. [4413], Cella shows CIO digital twin may work in connection with the EMP 60000 to provide simulations, predictions, statistical summaries, decision support based on analytics, machine learning, and/or other AI and learning-type processing of inputs (e.g., performance data, sensor data, and the like));
and implement the one or more software code related operations performed in the digital twin in the software code environments (Para. [4398] a CTO digital twin 60630 may be configured to allow a user to research, create, track and report on a technology, development, and/or technology or engineering department initiative including, but not limited to, a new product development, update, enhancement, replacement, upgrade, or the like).
and implement the one or more software code related operations performed in the digital twin in the software code environments (Para. [4398], Cella shows a CTO digital twin 60630 may be configured to allow a user to research, create, track and report on a technology, development, and/or technology or engineering department initiative including, but not limited to, a new product development, update, enhancement, replacement, upgrade, or the like).
In addition, in the analogous art of issue identification and remediation, Goloubew teaches:
wherein the executable instructions cause the at least one processing device to: provide a digital twin of software code environments comprising: .. the software code production environment (Para. [33] Goloubew shows “using digital twin environment 136, the conditions of a production environment may be simulated so that code modules can be tested with regard to their ability to identify and remediate issues without concern for any adverse impacts to an actual production environment”);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Cella into the teachings of Sundaresan as modified to implement “wherein the executable instructions cause the at least one processing device to: provide a virtual environment to the one or more users to allow the one or more users to perform one or more software code related operations associated with the one or more entity applications”. The modification would have been obvious as one of ordinary skill in the art would be motivated to allow a user to identify events and consequences that occur as a result of multiple simultaneous or related faults or other events (Cella, Para. [4317]).
In addition, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate the teachings of Goloubew into the teachings of Sundaresan as modified to implement “wherein the executable instructions cause the at least one processing device to: provide a digital twin of software code environments comprising: .. the software code production environment”. The modification would have been obvious as one of ordinary skill in the art would be motivated Issues can be simulated in digital twin environment 136 to test for a code module's ability to identify the issues and subsequently remediate the issues (Goloubew, Para. [33]).
With regards to claim 12, it is a computer program product claim having similar limitations as cited in claim 7 above. Thus, claim 12 is also rejected under the same rationale as cited in the rejection of claim 7 above.
With regards to claim 18, it is a method claim having similar limitations as cited in claim 7 above. Thus, claim 18 is also rejected under the same rationale as cited in the rejection of claim 7 above.
Response to Arguments
Applicant's arguments regarding 35 U.S.C 101 filed 6/08/2026 have been fully considered and they are persuasive.
Applicant's arguments regarding 35 U.S.C 103 filed 6/02/2026 have been fully considered but they are not persuasive.
Applicant argues on Pg. 12 of 15, that Sundaresan in view of Singh does not disclose the encoder and decoder layer architecture. However, it is shown that the encoder and decoder layer architecture is known in the art as Sundaresan shows (Fig. 3A-B; Para. [48], “The aggregated decoder output is input to each output layer 235, 237. Edit action output layer 235 includes a linear layer 238 and a SoftMax layer 240.”); and Doshi shows (Section 1: Architecture Overview [First Image]; Doshi shows in the image a transformer architecture common in the art for the purpose of education. Included is a decoder comprising of a positional encoding layer. Section 5: Decoder [Para. 4-6]; Doshi shows “Each of these sub-layers, Self-attention, Encoder-Decoder attention, and Feed-forward, have a residual skip-connection around them, followed by a Layer-Normalization.”). It would be obvious for Singh to use the encoder and decoder layer architecture as Singh teaches using “other types of sequence-to-sequence models and/or encoder-decoders” for translation.
Applicant further argues on Pg. 13, that Sobran uses a Siamese neural network for “for defect likelihood detection in source code commits based on log analysis-not for analyzing translated code to identify patterns associated with known merge conflicts”. Examiner notes merge conflicts can only happen when committed code conflicts with existing program structure. Sobran teaches analyzing source code commits and detecting patterns in commits that could cause conflicts using logs comparing historical commits to the new commit to give a probability threshold to identify defects (conflicts).
Applicant further argues on Pg. 13, “none of the cited references teach or suggest the unified build and packaging limitation now recited in the amended independent claims, which requires that build and packaging is performed ‘using build and packaging tools and requirements associated with the first software programming language, wherein the build and packaging of the unified code does not require multiple build and packaging tools associated with the one or more software programming languages.’ ”. Examiner notes code that has been translated is built and packaged as an application the same way as code that has been developed natively in a software language. The translated code requires the same building and packaging tools as an application that is all built in one language. Merely requiring the tools to build and package for one language does not change that using building and packaging an application with multiple languages is still capable of building and packaging a single software language.
The 103 rejection for claims 8 and 15 is maintained as they recite similar limitations to claim 1 above and the rejection to claim 1 is maintained as well.
For these reasons above, the examiner finds these arguments unpersuasive and maintains that the rejection under 35 U.S.C. 103 is proper.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US 20260178676 A1 – This prior art shows a transformer sequence to sequence model architecture
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.
/Z.A.M./Examiner, Art Unit 2193
/Chat C Do/Supervisory Patent Examiner, Art Unit 2193
Prosecution Insights