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
Amendments and Arguments filed on 3/11/26 have been entered.
With this office action, the examiner maintains the 103 rejections based on the amended claims.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 8, 15 are rejected under 35 U.S.C. 103 as being unpatentable over Nori (US 20250148220 A1) in further view of Gardner (US 12008332 B1), Miller (US 20210374186 A1 ) and Aggarwal (US 20250094732 A1)
With respect to claims 1, 8 and 15, Nori teaches
(claim 1) A method of saving prompt text length for large language models, executable by a processor, comprising ([0135]For example, a presentation component such as a display device is an I/O component in some examples, and some examples of processors have their own memory. Distinction is not made between such categories as “workstation,” “server,” “laptop,” “hand-held device,” etc., as all are contemplated within the scope of FIG. 6 and the references herein to a “computing device.” Memory 1712 may take the form of the computer storage media referenced below and operatively provide storage of computer-readable instructions, data structures, program modules and other data for the computing device 1700. In some examples, memory 1712 stores one or more of an operating system, a universal application platform, or other program modules and program data. Memory 1712 is thus able to store and access data 1712a and instructions 1712b that are executable by processor 1714 and configured to carry out the various operations disclosed herein):
(claim 8) A computer system for saving prompt text length for large language models, the computer system comprising: one or more computer-readable storage media configured to store computer program code ([0135]For example, a presentation component such as a display device is an I/O component in some examples, and some examples of processors have their own memory. Distinction is not made between such categories as “workstation,” “server,” “laptop,” “hand-held device,” etc., as all are contemplated within the scope of FIG. 6 and the references herein to a “computing device.” Memory 1712 may take the form of the computer storage media referenced below and operatively provide storage of computer-readable instructions, data structures, program modules and other data for the computing device 1700. In some examples, memory 1712 stores one or more of an operating system, a universal application platform, or other program modules and program data. Memory 1712 is thus able to store and access data 1712a and instructions 1712b that are executable by processor 1714 and configured to carry out the various operations disclosed herein); and
(claim 8) one or more computer processors configured to access said computer program code and operate as instructed by said computer program code, said computer program code including ([0135]For example, a presentation component such as a display device is an I/O component in some examples, and some examples of processors have their own memory. Distinction is not made between such categories as “workstation,” “server,” “laptop,” “hand-held device,” etc., as all are contemplated within the scope of FIG. 6 and the references herein to a “computing device.” Memory 1712 may take the form of the computer storage media referenced below and operatively provide storage of computer-readable instructions, data structures, program modules and other data for the computing device 1700. In some examples, memory 1712 stores one or more of an operating system, a universal application platform, or other program modules and program data. Memory 1712 is thus able to store and access data 1712a and instructions 1712b that are executable by processor 1714 and configured to carry out the various operations disclosed herein):
(claim 16) A computer program product for saving prompt text length for large language models, comprising: one or more computer-readable storage devices ; and
program instructions stored on at least one of the one or more computer-readable storage devices, the program instructions configured to cause one or more computer processors to:
prompting a large language model with a first prompt to receive a first result (Nori ¶ [0005] Example solutions for processing LLM prompts include: receiving an input large language model (LLM) prompt [first prompt]; creating a first LLM prompt based on the input LLM prompt, the first LLM prompt representing a first step toward generating a first solution to the input LLM prompt; submitting the first LLM prompt to an LLM as a first sub-query, thereby resulting in the generation of a first LLM output [first result]; creating a second LLM prompt [second prompt] based on the input LLM prompt, the second LLM prompt representing a second step toward generating the first solution to the input LLM prompt, the second LLM prompt including the first LLM output; submitting the second LLM prompt to the LLM as a second sub-query, thereby resulting in the generation of a second LLM output [second result]; and transmitting the second LLM output as the solution to the input LLM prompt.);
prompting the large language model with [[ the generated summary]] to receive a second result (Nori ¶ [0005] Example solutions for processing LLM prompts include: receiving an input large language model (LLM) prompt [first prompt]; creating a first LLM prompt based on the input LLM prompt, the first LLM prompt representing a first step toward generating a first solution to the input LLM prompt; submitting the first LLM prompt to an LLM as a first sub-query, thereby resulting in the generation of a first LLM output [first result]; creating a second LLM prompt [second prompt] based on the input LLM prompt, the second LLM prompt representing a second step toward generating the first solution to the input LLM prompt, the second LLM prompt including the first LLM output; submitting the second LLM prompt to the LLM as a second sub-query, thereby resulting in the generation of a second LLM output [second result]; and transmitting the second LLM output as the solution to the input LLM prompt.);
generating a text output associated with the first prompt based on prompting the large language model with a second prompt [[generated by the trained summary model.]] ([0005] Example solutions for processing LLM prompts include: receiving an input large language model (LLM) prompt [first prompt]; creating a first LLM prompt based on the input LLM prompt, the first LLM prompt representing a first step toward generating a first solution to the input LLM prompt; submitting the first LLM prompt to an LLM as a first sub-query, thereby resulting in the generation of a first LLM output [first result]; creating a second LLM prompt [second prompt] based on the input LLM prompt, the second LLM prompt representing a second step toward generating the first solution to the input LLM prompt, the second LLM prompt including the first LLM output; submitting the second LLM prompt to the LLM as a second sub-query, thereby resulting in the generation of a second LLM output [second result]; and transmitting the second LLM output as the solution to the input LLM prompt.)
Nori does not explicitly disclose however Gardner teaches generating a summary based on prompting a summary model with the first prompt (Gardner¶ Col7ll59-65 The prompts and APIs abstract the underlying model implementations, allowing the system to orchestrate and leverage an evolving ensemble of AI technologies as new models emerge. The system's model-agnostic approach facilitates experimentation and optimization to determine the best combination of models and prompts for each summarization task.)
training the summary model based on reinforcement learning, wherein the reinforcement learning comprises calculating a score associated with an output of the summary model (Gardner ¶ Col15ll31-33 In addition to supervised learning on human prompts, reinforcement learning (RL) may be used to further enhance prompt engineering, ¶ Col18ll50-61 Programmatically generating a wide range of prompts for the LLM with varying instructions, abstraction levels, word limits, and other attributes; Analyzing the LLM's output summaries to measure attributes like length, use of inferences, integration of external data, and conciseness; Scoring summary quality through metrics like information coverage, coherence, readability, and human validation; and/or Using reinforcement learning to associate prompt attributes with summary outcomes to determine optimal prompts)
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori to include the summary model of Gardner in order to reduce information quantity while retaining quality, (Col2ll24-26, Gardner);
None of Nori and Gardner explicitly disclose however Miller teaches wherein the score is calculated based on dividing a logarithm of the similarity score of the first result and the second result by a [[maximum number of tokens associated with the large language model]] (Miller ¶[0076] For example, when determining a features score (e.g., similarity score) for the “Release Year” label, the database management application may determine an absolute value of the difference in dates (e.g., in days and/or years) [similarity=difference], take the logarithm of such absolute value, and normalize the logarithm of the absolute value). Examiner Note: days are mapped to first/second value and their difference indicates similarity. The result is then divided (normalized).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori in view of summary model of Gardner to include scoring of Miller in order to generate an efficient summary for post-processing;
None of Nori, Gardner, Kim and Bhagwan explicitly disclose however, Aggarwal teaches maximum number of tokens associated with the large language model (Aggarwal ¶[0094] The LLM then selects the summary that has the highest cumulative token probability as the summary for the content received in 202. In certain examples, the cumulative token probability is calculated by summing the log probabilities of each token in a sequence. To select a summary based on token probabilities, in certain examples, the LLM uses the average log probability (by summing the log probabilities and dividing by the number of tokens in the summary) to avoid length bias.)
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori in view of summary model of Gardner in view of similarity of Kim in view of log of Bhagwan to include maximum tokens of Aggarwal in order to improve length bias;
Claims 5, 12 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Nori, Gardner Miller and Aggarwal in further view of McDonald (US 20250168186 A1).
With respect to claims 5, 12 and 19, none of Nori, Gardner Miller and Aggarwal explicitly disclose however McDonald teaches wherein background data and a prompt template associated with the first prompt are refined based on a task description associated with the prompt (McDonald ¶[0079] With reference to FIG. 1B and FIG. 1E, a host machine workflow 100E, associated with a host machine (e.g., host machine 116) and a summarizer (e.g., summarize 120C) is used to update a context [background data] associated with a prompt template).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori in view of summary model of Gardner in view of scoring of Miller to include background data of McDonald in order to enable efficiency and customization.
Claims 6, 7, 12, 14 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Nori, Gardner Miller and Aggarwal in further view of Nishimoto (US 20250173630 A1).
With respect to claims 6, 12 and 20, none of Nori, Gardner Miller and Aggarwal explicitly disclose however Nishimoto teaches wherein the large language model comprises a transformer architecture (Nishimoto ¶ [0124] The learning method of the language model is not particularly limited, but as an example, the language model may be learned to output at least one sentence that includes the input character string. As a specific example, the language model is a GPT (Generative Pre-Transformer) that outputs a sentence including the input character string by predicting a character string that has a high probability of following the input character string.).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori in view of summary model of Gardner in view of scoring of Miller to include transformer of Nishimoto in order to enable efficiency implementation of LLM.
With respect to claims 7 and 14, Nishimoto further teaches wherein the transformer architecture corresponds to a generative pre-transformer(Nishimoto ¶ [0124] The learning method of the language model is not particularly limited, but as an example, the language model may be learned to output at least one sentence that includes the input character string. As a specific example, the language model is a GPT (Generative Pre-Transformer) that outputs a sentence including the input character string by predicting a character string that has a high probability of following the input character string.).
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori in view of summary model of Gardner in view of scoring of Miller to include transformer of Nishimoto in order to enable efficiency implementation of LLM.
Claims 21,22, and 23 are rejected under 35 U.S.C. 103 as being unpatentable over Nori, Gardner Miller and Aggarwal in further view of Fayyaz (US 20250086187 A1).
With respect to claims 21, 22 and 23 none of Nori, Gardner Miller and Aggarwal explicitly disclose however Nishimoto teaches wherein the second prompt has a shorter length than the first prompt ([0029] In some implementations, the main-system model 112 includes a base model 116 and a prompt-compressing model 118. In some implementations, for example, the base model 116 is a large language model, e.g., having the model architecture described in Section B. The base model 116 transforms a task description received by the client device 104 into an initial task prompt having a first size. In some implementations, the prompt-compressing model 118 includes a linear fully-connected feed-forward neural network (FFN) having one or more layers. The prompt-compressing model 118 is trained to convert the initial task prompt into a final task prompt having a second size that is smaller than the first size. In other words, the prompt-compressing model 118 compresses the initial task prompt.)
It would have been obvious to one of ordinary skill in the art prior to the effective filing date of the invention to modify the LLM of Nori in view of summary model of Gardner in view of scoring of Miller to include transformer of Nishimoto in order to increase accuracy of model responses (Fayyaz [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 extension fee 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 date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ATHAR N PASHA whose telephone number is (408)918-7675. The examiner can normally be reached Monday-Thursday Alternate Fridays, 7:30-4:30 PT.
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, Daniel Washburn can be reached on (571)272-5551. 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.
/ATHAR N PASHA/Primary Examiner, Art Unit 2657
Prosecution Insights