Prosecution Insights
Last updated: November 04, 2026
Application No. 18/539,067

SYSTEM AND METHOD OF LLM TRAINING FOR NETWORK STATE CONVERSATIONAL ANALYTICS

Non-Final OA §103
Filed
Dec 13, 2023
Priority
Nov 08, 2023 — CIP of 18/504,991
Examiner
LE, JESSICA N
Art Unit
2169
Tech Center
2100 — Computer Architecture & Software
Assignee
Aviz Networks, Inc.
OA Round
3 (Non-Final)
73%
Grant Probability
Favorable
3-4
OA Rounds
10m
Est. Remaining
99%
With Interview

Examiner Intelligence

Grants 73% — above average
73%
Career Allowance Rate
373 granted / 514 resolved
+17.6% vs TC avg
Strong +28% interview lift
Without
With
+28.1%
Interview Lift
resolved cases with interview
Typical timeline
3y 9m
Avg Prosecution
12 currently pending
Career history
531
Total Applications
across all art units

Statute-Specific Performance

§101
17.9%
-22.1% vs TC avg
§103
51.4%
+11.4% vs TC avg
§102
13.3%
-26.7% vs TC avg
§112
12.4%
-27.6% vs TC avg
Black line = Tech Center average estimate • Based on career data from 514 resolved cases

Office Action

§103
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . DETAILED ACTION Continued Examination Under 37 CFR 1.114 A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 04/21/2026 has been entered. Continuity The instant application is claimed continue-in-part (CIP) of application number 18/504,991 filed on 11/08/2023. Status of Claims Claims 1, 6, and 11 are independent claims, and are amended. Claims 1-15 are pending in this application. Information Disclosure Statement The information disclosure statement (IDS) submitted on 05/30/2026 has been considered and recorded. The submission is in compliance with the provisions of 37 CFR §1.97. See form PTO-1449 singed and attached hereto. Response to Arguments Referring claim rejections under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, the rejections have been withdrawn in view of amendment to the claims. Referring claim rejections under 35 U.S.C. 103, Applicant's arguments (Remarks, pages 6-8) with respect to the amended limitations in the respective independent claim 1 (same as to claims 6 and 11) have been considered but are moot in view of the new grounds of rejection necessitated by applicant's amendment to the claims. Applicant's newly amended features are taught implicitly, expressly, or impliedly by the prior art of record. Please see the rejections set forth below for details. 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. Claims 1, 4, 6, 8, 11, and 13 are rejected under 35 U.S.C. 103 as being unpatentable over He et al., US Pub. No. 2021/0319357 A1 (hereinafter as “He”) in view of Nguyen et al., US Patent No. 10,546,001 (hereinafter as “Nguyen”), and further in view of MENGKE et al., US Pub. No. 2025/0148400 A1 (hereinafter as “Mengke”). Regarding claim 1, He teaches: a method, comprising: receiving, by one or more agent applications, a query from a user (see par. [0004] wherein the machine learning (ML) applications is interpreted as the agent application(s), pars. [0039] “detected user input that are provided as inputs to the data preparation recommendation model”, and [0050] e.g., “a query”); providing a network state dataframe to the one or more agent applications (par. [0053] “Pandas API calls” which is/are implemented as the network_state dataframe, see further in pars. [0068-70] e.g., “Pandas data-frames”); appending a predetermined number of initial entries from the network state dataframe to a suffix of a prompt to serve as contextual information (Fig. 4 as shown the pandas’ code snippet being the contextual information which is served as suffix of a prompt; see par. [0050] e.g., “performing a query using the file suffix (.ipynb)”, par. [0069] “seven Pandas API data preparation operation calls that either (i) take data-frame objects as parameters or (ii) produce the data-frames concat, dropna, fillna, groupby, melt, merge, or pivot”, wherein the “seven Pandas API data preparation operation calls” to take dataframe objects as parameters is interpreted as the predetermined number of initial entries; see further in pars. [0164] “contextual information”, and [0173] “utilize the data preparation recommendation model to generate data preparation recommendation prompt comprising at least one of (i) an instantiation of a specific operator or (ii) a next operator”, and [0180]; and par. [0182], wherein the threshold is implemented as inherited to the appended predetermined number of initial entries to a data preparation recommendation prompt being displayed to a user one the probability of the data preparation recommendation model output in limitation). He further teaches the “generating a prompt based on the data preparation recommendation model output” (see par. [0191]) using the well-known Pandas network_state dataframes including the code snippet, syntax, parameters, etc. (see pars. [0068-71]). However, He does not explicitly teach: “constructing a standardized prompt template, wherein the query is embedded within the standardized prompt template.” In the same field of endeavor (i.e., data processing), Nguyen teaches: “constructing a standardized prompt template, wherein the query is embedded within the standardized prompt template.” (Fig. 9 as shown the prompt template, wherein the query/queries embedded in template, and Col. 25, lines 50-58; and further in Col. 28, lines 43-67, and Col. 29, lines 1-20: “…The natural language query processor 310 supports query template that match queries that request statistics of a particular attribute of the dataset. For example, the natural language query processor 310 supports queries of the form “what is the standard deviation of X?” The query template for such queries comprises the following components: (1) the phrase “what is the”, (2) a phrase pattern of a statistical operation belonging from a predefined list, for example, “max”, “min”, “average”, “median”, “standard deviation”, “25th-75th quartiles”, or “distribution.” (3) the keyword “of”, and (4) an attribute (or user defined metric.) The query template allows an optional extension of a filter clause) Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the instant application to combine the teachings of the cited references because the teachings of Nguyen including the standard prompt template embedded the query would have provided He with the teachings of using Pandas dataframes to implement the suffix of a prompt exploits contextual information in the natural language processing neural language modes for utilizing a data preparation recommendation model, and a recurrent neural network (RNN) to create a machine learned predictive model for this prediction task more efficient (Nguyen: Figs. 2 and 4; and pars. [0068-71, 164, 173, and 182]). He and Nguyen, in combination, teach code snippet (He: Fig. 4; and par. [0071]), machine learning language and natural language query templates (He: pars. [0004 and 164]; Nguyen: Abstract). However, He and Nguyen do not explicitly teach: “configured to instruct one or more Large Language Models (LLMs) to generate executable code snippet based on the contextual information for analyzing the network state dataframe to resolve the user’s query;” “channeling the standardized prompt template to one or more Large Language Models (LLMs); utilizing a Generative Pre-trained Transformer (GPT) Application Programming Interface (API) to generate a generated code snippet; executing the generated code snippet on the network state dataframe to create a resulting data analysis output; and relaying the resulting data analysis output from the execution of the generated code snippet back to the user as a direct response to the query.” In the same field of endeavor (i.e., data processing), Mengke teaches: configured to instruct one or more Large Language Models (LLMs) to generate executable code snippet based on the contextual information for analyzing the network state dataframe to resolve the user’s query; (see Fig. 1, element 103 as Large Language Model (LLM); Fig. 3, element 308; and Abstract: “generate a prompt for a Large Language Model (LLM) comprising the user activity data and instructing the LLM to generate a report based on the user activity data, and submit the prompt to the LLM and receive the report generated by the LLM”); and channeling the standardized prompt template to one or more Large Language Models (LLMs) (see Fig. 3, element 310, e.g., submits prompt to LLM (e.g., GPT model); and Abstract: “…submit the prompt to the LLM and receive the report generated by the LLM”; par. [0099] “a prompt template for the prompt to the LLM, wherein the API is to access the prompt template to generate the prompt for the LLM”); utilizing a Generative Pre-trained Transformer (GPT) Application Programming Interface (API) to generate a generated code snippet (see Fig. 5, element 152 as shown the GPT API with the code snippet is generated; par. [0036] “the API submits the prompt to the LLM 310, as described above. In various examples, the LLM may be a GPT model.”; par. [0060] “The grid service 153 will also include a cloud provisioning service 153 to deploy or utilize a GPT model on the cloud computing platform 156. The grid service 153 thus securely handles the prompt to the GPT model and the response from the GPT model back to the browser 150 of the administrator.”; and par. [0057], e.g., “a summary icon or similarly labeled button to initiate the AI assistant. This invokes an administrator portal 141 and the API configured to obtain an LLM or GPT summary…”); executing the generated code snippet on the network state dataframe to create a resulting data analysis output (see Fig. 5 as shown the data frame, Fig. 6 as shown the resulting data analysis output; and pars. [0049-51] teach data analysis, and [0076] “the frameworks 718 (also sometimes referred to as middleware) provide a higher-level common infrastructure that may be used by the applications 720 and/or other software modules. For example, the frame works 718 may provide various graphic user interface (GUI) functions, high-level resource management, or high-level location services. The frameworks 718 may provide a broad spectrum of other APIs for applications 720 and/or other software modules”; par. [0077], wherein the framework 718 embeds the API dataframe); and relaying the resulting data analysis output from the execution of the generated code snippet back to the user as a direct response to the query (see Figs. 3 and 5; and par. [0032] e.g., “The result is a report in real time that provides the administrator with an accurate view of what is happening on the collaboration system and what actions might be needed to better administer the system. Alternatively, collection of the activity data and the prompt to the LLM or a summarization report could be performed automatically, for example, on a periodic basis. In such an example the resulting report could be offered to the administrator when available without the administrator having to initiate the AI assistant feature to produce the report as described herein”; and further in pars. [0035 and 73]). Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the instant application to combine the teachings of the cited references because the teachings of Mengke would have provided He and Nguyen with the above indicated limitations for allowing a skill artisan in motivation in constructing and using the prompt script template to generate the code snippet via GPT API to create the resulting data analysis output to user (Mengke: Figs. 1-4; Abstract, and pars. [0029, 35 and 57]). Regarding claim 2, He teaches: “specifying one or more pandas commands to execute in order to formulate an appropriate response to the query.” (see Fig. 4, and pars. [0068-69] via “seven Pandas API data preparation operation calls that either (i) take data-frame objects as parameters or (ii) produce the data-frames concat, dropna, fillna, groupby, melt, merge, or pivot”) Regarding claim 4, He and Mengke, in combination, teach: “receiving processed data in CSV format.” (He: par. [0071] e.g., “two CSV files”; and Mengke: par. [0054] “the data provided to the LLM needs to be prepared in a way that is compatible with the model. This includes being aware of the token limit and converting the data into a format like JSON or CSV to string.”) Claims 6-8, 11, and 13 are rejected in the analysis of above claims 1-2 and 4; and therefore, the claims are rejected on that basis. Claims 3, 9, and 14 are rejected under 35 U.S.C. 103 as being unpatentable over He, Nguyen and Mengke, and further in view of Lee et al., US Pub. No. 2025/0147999 A1 (hereinafter as “Lee”). Regarding claim 3, the claim is rejected by the same reasons set forth above to claim 1. However, He, Nguyen, and Mengke do not explicitly teach: “wherein the relaying is done post-execution.” Regarding claim 3, Lee teaches: “wherein the relaying is done post-execution.” (par. [0045] “The resulting sequence of output tokens 216 can then be converted to a text sequence in post-processing. For example, each output token 216 can be an integer number that corresponds to a vocabulary index. By looking up the text segment using the vocabulary index, the text segment corresponding to each output token 216 can be retrieved, the text segments can be concatenated together, and the final output text sequence can be obtained”, wherein the “post-processing” is interpreted as post-execution) Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the instant application to combine the teachings of the cited references because the teachings of Lee would have provided He, Nguyen, and Mengke with the above indicated limitation for allowing a skill artisan in motivation to perform the done post processing/execution for final output/result to user application(s) (Lee: Figs. 5-7 and par. [0045]). Claims 9 and 14 are rejected in the analysis of above claim 3; and therefore, the claims are rejected on that basis. Claims 5, 10, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over He, Nguyen, Mengke, and Lee, and further in view of MORRIS, US Pub. No. 2025/0111143 A1 (hereinafter as “Morris”). Regarding claim 5, the claim is rejected by the same reasons set forth above to claim 1 and 3. Furthermore, Mengke teaches: “the temperature” (par. [0055] “In some LLMs, this parameter is referred to as the temperature. Consequently, the LLM prompt 131 may be accompanied by a temperature setting of about 0.2 and no more than 0.4. This will ensure that the generated insights are accurate and relevant.”). However, He, Nguyen, Mengke, and Lee do not explicitly teach that “setting a temperature of the one or more large language models to zero.” In the same field of endeavor (i.e., data processing), Morris teaches: “setting a temperature of the one or more large language models to zero” (see par. [0042] “the response generation module 102 may provide a request to a large language model 104 to ask whether the prompt relates to contemporaneous data. Using ChatGPT as an example of the LLM, this may be implemented using a prompt which directly requests whether the query relates to contemporaneous data. This request may utilise a zero-temperature setting within ChatGPT which asks for a purely deterministic yes/no answer. LLM 104 may also be a private LLM”; and further in par. [0069] e.g., “uses temperature setting of 0, i.e. zero”). Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the instant application to combine the teachings of the cited references because the teachings of Morris would have provided He, Nguyen, Mengke, and Lee with the above indicated limitation(s) for allowing a skill artisan in motivation to perform question and answer within ChatGPT data processing using large language models (LLMs) more efficient. Claims 10 and 15 are rejected in the analysis of above claim 5; and therefore, the claims are rejected on that basis. Prior Arts The prior art made of record on form PTO-892 and not relied upon is considered pertinent to applicant's disclosure. Applicant is required under 37 C.F.R. § 1.111(c) to consider these references fully when responding to this action. It is noted that any citation to specific, pages, columns, lines, or figures in the prior art references and any interpretation of the references should not be considered to be limiting in any way. A reference is relevant for all it contains and may be relied upon for all that it would have reasonably suggested to one having ordinary skill in the art. See In re Heck, 699 F.2d 1331, 1332-33, 216 USPQ 1038, 1039 (Fed. Cir. 1983) (quoting In re Lemelson, 397 F.2d 1006, 1009, 158 USPQ 275,277 (CCPA 1968)); Merck & Co. v. Biocraft Laboratories, 874 F.2d 804, 10 USPQ2d 1843 (Fed. Cir.), cert. denied, 493 U.S. 975 (1989). Conclusion Any inquiry concerning this communication or earlier communications from the examiner should be directed to Jessica N. Le whose telephone number is (571)270-1009. The examiner can normally be reached M-F 9:30 am - 5:30 pm (EST). Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, SHERIEF BADAWI can be reached at (571) 272-9782. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000. /Jessica N Le/Examiner, Art Unit 2169 /SHERIEF BADAWI/Supervisory Patent Examiner, Art Unit 2169
Read full office action

Prosecution Timeline

Dec 13, 2023
Application Filed
Jun 25, 2025
Non-Final Rejection mailed — §103
Sep 25, 2025
Response Filed
Oct 21, 2025
Final Rejection mailed — §103
Apr 21, 2026
Request for Continued Examination
Apr 26, 2026
Response after Non-Final Action
Aug 21, 2026
Non-Final Rejection mailed — §103 (current)

Precedent Cases

Applications granted by this same examiner with similar technology

Patent 12769755
SYSTEMS AND METHODS FOR CREATING GENERATIVE AI FRAMEWORKS ON NETWORK STATE TELEMETRY
2y 11m to grant Granted Oct 20, 2026
Patent 12737348
DATABASE SYSTEM PUBLIC TRUST LEDGER ARCHITECTURE
5y 0m to grant Granted Sep 15, 2026
Patent 12724927
ENHANCED BLOCKCHAIN DATA COMPUTING PLATFORM FOR STRATEGIC MASTER DATA MANAGEMENT
3y 6m to grant Granted Sep 01, 2026
Patent 12717772
ASSESSMENT NODE AND TOKEN ASSESSMENT CONTAINER
5y 6m to grant Granted Aug 25, 2026
Patent 12705268
APPARATUS, METHOD, AND COMPUTER PROGRAM PRODUCT FOR DISTRIBUTED SEMANTIC SEARCH
3y 4m to grant Granted Aug 11, 2026
Study what changed to get past this examiner. Based on 5 most recent grants.

Strategy Recommendation AI-generated — please review before filing

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

Prosecution Projections

3-4
Expected OA Rounds
73%
Grant Probability
99%
With Interview (+28.1%)
3y 9m (~10m remaining)
Median Time to Grant
High
PTA Risk
Based on 514 resolved cases by this examiner. Grant probability derived from career allowance rate.

Sign in with your work email

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

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

Free tier: 3 strategy analyses per month