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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 04/14/2026 has been entered.
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 .
Examiner’s Comment
This Action is in response to the Request for Continued Examination filed on 04/14/2026 with Amended Claims and Applicant's Remarks filed on 03/12/2026.
Applicant has amended claims 1, 10, and 11 according to Amendments filed on 03/12/2026. Claims 1-4 and 7-11 are pending and currently under consideration for patentability.
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 1-4 and 7-11 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claims are directed to a judicial exception (i.e., a law of nature, natural phenomenon, or abstract idea) without significantly more.
Step 1: In a test for patent subject matter eligibility, claims 1-4 and 7-11 are found to be in accordance with Step 1 (see 2019 Revised Patent Subject Matter Eligibility), as they are related to a process, machine, manufacture, or composition of matter. Claims 1-4, 7-9 recite a system, claim 10 recites a method, and claim 11 recites a non-transitory computer-readable medium. When assessed under Step 2A, Prong I, they are found to be directed towards an abstract idea. The rationale for this finding is explained below:
Step 2A, Prong I: Under Step 2A, Prong I, claims 1-4 and 7-11 are directed to an abstract idea without significantly more, as they all recite a judicial exception. Independent Claims 1, 10, and 11 recite limitations directed to the abstract idea including “acquiring a prompt that includes pieces of information on a plurality of items; acquiring real-time evaluation information that is information indicating evaluation on a service using generation information; and estimating degrees of importance of the plurality of items based on the real-time evaluation information that is acquired; automatically generating a new prompt in real-time from one or more of the plurality of items based on the degrees of importance; and inputting the new prompt to generate the new generation information corresponding to the new prompt”. These further limitations are not seen as any more than the judicial exception. Claims 1, 10, and 11 recite additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. Generating a new prompt based on degrees of importance in order to generate new generation information is considered to be an abstract idea, specifically, certain methods of organizing human activity; such as managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions) because the claims are directed to managing a relationship between parties (i.e. evaluating a service) in order to estimate information (i.e. importance of items). Furthermore, generating a new prompt based on degrees of importance in order to generate new generation information is also considered to fall under another grouping of abstract idea, specifically, Mental Processes; such as concepts performed in the human mind (including an observation, evaluation, judgment, opinion) because the claims are directed to acquiring information (i.e. prompt and real-time evaluation of a service); estimating information (i.e. degrees of importance of items) based on the acquired information; generating a new prompt based on degrees of importance; and inputting the new prompt to generate new information corresponding to the new prompt which can all be performed in the human mind. Finally, generating a new prompt based on degrees of importance in order to generate new generation information is also considered to fall under another grouping of abstract idea, specifically, Mathematical Concepts; such as mathematical relationships, formulas or equations, and calculations, because the claims are directed to estimating degrees of importance of items based on an evaluation on a service which may require mathematical relationship in order to determine the estimation. Therefore, under Step 2A, Prong I, Claims 1, 10, and 11 are directed towards an abstract idea.
Step 2A, Prong II: Step 2A, Prong II is to determine whether any claim recites any additional element that integrate the judicial exception (abstract idea) into a practical application. Claims 1, 10, and 11 recite additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. The additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt” are not found to integrate the judicial exception into a practical application because they are seen as adding the words “apply it” (or an equivalent such as “using”) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) and generally linking the use of the judicial exception to a particular technological environment or field of use – see MPEP 2106.05(h). Here, the “generative AI comprises a large language model or an image generative AI” is merely being used to describe the text or image or information that is generated and the “machine learning that uses regression analysis, gradient boosting, or neural networks” is merely being used to as a tool to estimate the degrees of importance of items. Accordingly, alone, and in combination, these additional elements are seen as using a computer or tool to perform an abstract idea, adding insignificant-extra-solution activity to the judicial exception. They do no more than link the judicial exception to a particular technological environment or field of use, i.e. generative AI or machine learning, and therefore do not integrate the abstract idea into a practical application. The courts decided that although the additional elements did limit the use of the abstract idea, the court explained that this type of limitation merely confines the use of the abstract idea to a particular technological environment and this fails to add an inventive concept to the claims (See Affinity Labs of Texas v. DirecTV, LLC,). Under Step 2A, Prong II, these claims remain directed towards an abstract idea.
Step 2B: Claims 1, 10, and 11 recite additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. These additional limitations do not integrate the judicial exception (abstract idea) into a practical application because of the analysis provided in Step 2A, Prong II. The independent claims do not include additional elements or a combination of elements that result in the claims amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements listed amount to no more than mere instructions to apply an exception using a generic computer component. In addition, the applicant’s specifications describe “general purpose graphic processing unit”, pages 28-31, for implementing the computer, which do not amount to significantly more than the abstract idea of itself, which is not enough to transform an abstract idea into eligible subject matter. Furthermore, there is no improvement in the functioning of the computer or technological field, and there is no transformation of subject matter into a different state. Under Step 2B in a test for patent subject matter eligibility, these claims are not patent eligible.
Dependent claims 2-4 and 7-9 further recites independent claim 1. Dependent claims 2-4 and 7-9 when analyzed as a whole are held to be patent ineligible under 35 U.S.C. 101 because the additional recited limitation fail to establish that the claims are not directed to an abstract idea:
Under Step 2A, Prong I, these additional claims only further narrow the abstract idea set forth in Claims 1, 10, and 11. For example, claims 2-4 and 7-9 describe the limitations for generating a new prompt based on degrees of importance in order to generate new generation information – which is only further narrowing the scope of the abstract idea recited in the independent claims.
Under Step 2A, Prong II, for dependent claims 2-4 and 7-9, there are no additional elements introduced. Thus, they do not present integration into a practical application, or amount to significantly more.
Under Step 2B, the dependent claims do not include any additional elements that are sufficient to amount to significantly more than the judicial exception. Additionally, there is no improvement in the functioning of the computer or technological field, and there is no transformation of subject matter into a different state. As discussed above with respect to integration of the abstract idea into a practical application, the additional claims do not provide any additional elements that would amount to significantly more than the judicial exception. Under Step 2B, these claims are not patent eligible.
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.
Claim(s) 1-4 and 7-11 are rejected under 35 U.S.C. 103 as being unpatentable over U.S. Patent 10,909,604 to Zappella in view of U.S. Publication 2025/0168056 to Lasso.
With respect to Claim 1:
Zappella teaches:
An information processing apparatus comprising: a processor, the processor is configured to (Zappella: Col. 18 Lines 47-60):
acquire a prompt that includes pieces of information on a plurality of items (i.e. acquire input that includes features of content on items) (Zappella: Col. 8 Lines 25-34 “Corresponding to individual ones of the ICEs provided by the content sources and the kinds of machine learning models (such as bandit models) to be used for optimization, one or more feature generators 123 of the ICOS may generate respective sets of input features in the depicted embodiment. The feature sets may, for example, include vectors of numerical elements which collectively represent the text of a book description, vectors which represent various characteristics of a book cover image, and so on.”);
acquire real-time evaluation information that is information indicating evaluation on a service using generation information that is information that is generated by using generative Al based on the prompt (i.e. acquiring effectiveness information on a service or items, wherein the effectiveness is determined using artificial intelligence on a prompt that includes information of the items) (Zappella: Cols. 2-3 Lines 60-15 “In at least some embodiments, one or more of the techniques described herein may be employed at a largescale content provider such as an e-retailer which has an inventory comprising millions of items, where the producers of at least some of the items may have relatively limited expertise regarding the effectiveness of different types of content elements with respect to increasing sales or other metrics of interest, and where at least some types of informational content that can be provided for such items is somewhat subjective. Such item producers may be referred to as the "long tail" of the distribution of item producers in some embodiments. Although such item producers may have limited experience and/or resources with respect to promoting their own items, the e-retailing organization ( or some affiliated organization) may in some embodiments have access to substantial artificial intelligence or machine learning resources, as well as an extensive collection of data regarding the effectiveness of various types if information that has been presented to potential consumers in the past. As such, various types of experimentation and optimization techniques may be employed at thee-retailer to help at least some such less-experienced item producers achieve their business objectives.” Furthermore, as cited in Col. 8 Lines 35-55 “In the depicted embodiment, a plurality of optimization iterations, which may also be referred to as ICE exploration iterations 147 (e.g., 147A or 147B) may be executed with respect to individual ones of the items 144 using a machine learning model in an online or continuous-learning mode. In a given iteration, a particular set of ICEs to be presented to a target audience may be identified, the identified set of ICEs may be presented, and metrics associated with the presentation may be collected and analyzed in various embodiments. Because the machine learning model is run in online mode, adjustments to the ICEs may be made relatively quickly based on the metrics and feedback collected in various embodiments. In at least some embodiments, if presentation constraints were indicated by the content sources, the ICEs to be presented may be selected according to those constraints. The metrics used to evaluate the success of a given set of presented ICEs may include, for example, a web link click metric, a sales metric, a cart insertion metric, a wish list insertion metric, or a session engagement length metric ( e.g., how long a potential consumer continues to interact with a music or video item or item collection).”); and
estimate degrees of importance of the plurality of items based on the real-time evaluation information using machine learning that uses [[regression analysis, gradient boosting,]] or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables (i.e. estimates/calculates metrics of importances or prioritization of items based on acquired information, wherein the item data is inputted or is the independent variable and the evaluation values are outputted or dependent on the item data) (Zappella: Col. 6 Lines 36-62 “Categories of ICEs which may be presented in one or more contexts may include, among others, still images, videos, audio recordings, text collections, web links, and so on. In some embodiments in which several different presentation contexts are available for a given item, the different contexts may be prioritized relative to one another-e.g., if far more consumers purchased an item as a result of viewing a recommendation than as a result of viewing an item details page, the optimization of the recommendation interface may be prioritized higher than the optimization of the details page. Records of earlier interactions performed for similar items (or the same item) in the different contexts may be analyzed to prioritize among contexts in such embodiments. In at least one embodiment, the ICOS may provide an indication of the particular presentation context or contexts for which a plurality of ICEs should be provided by a content source---e.g., it may be the case that a wider variety of ICES are needed for exploration with respect to an item details page than are needed for a search response context. In various embodiments, different types of effectiveness or utility metrics may be generated for ICEs at the machine learning models used during the iterative optimization phase. Such metrics may include, for example, web link click count metrics, sales metrics, shopping cart insertion metrics, wish list insertion metrics, and/or session engagement length metrics.” Furthermore, as cited in Col. 11 Lines 43-56 “The metrics may be analyzed, for example using a machine learning algorithm (such as a bandit algorithm 475) run in an online or continuous-learning mode, and the next selection 435 of the ICE variant set 410 may be based at least on part on the results of the metric analysis. In some embodiments, one or more additional machine learning models 476 (e.g., natural language understanding models) may be used to interpret the sentiment or semantics of some types of feedback such as reviews. In at least one embodiment, additional machine learning models 477 ( e.g., generational neural network models) may also or instead be employed to modify ICEs provided by the item owner, in accordance with specified constraints indicated by the item owner.”);
automatically generate a new prompt in real-time from one or more of the plurality of items based on the degrees of importance (i.e. determines information about items to be included in new iterations of effectiveness estimation/determination via input into machine learning model) (Zappella: Col. 17 Lines 38-66 “Features which can be used to represent, in the input provided to a machine learning model, various individual ICEs and the combinations of ICEs may be generated in the depicted embodiment (element 1107). The features may, for example, represent text descriptions as vectors within a word embedding space, images as encoded matrices of numeric values indicating colors and shapes, and so on… Depending on the effectiveness, new variants of ICE combinations may be selected for the next iteration in various embodiments. The iterations may be continued until a termination criterion (such a determination that the marginal improvement in effectiveness metrics with new exploratory iterations has fallen below a threshold) is met in some embodiments. For some items, such as items for which the rates of consumer interactions fall below a threshold, or the iterations may be conducted indefinitely.”); and
input the new prompt to the generative Al to cause the generative Al to generate the new generation information corresponding to the new prompt (i.e. new iterations that will be used as input for the machine learning model includes information on the items selected/determined to be of importance in order to generate new variants of the content) (Zappella: Col. 17 Lines 38-66 “Features which can be used to represent, in the input provided to a machine learning model, various individual ICEs and the combinations of ICEs may be generated in the depicted embodiment (element 1107). The features may, for example, represent text descriptions as vectors within a word embedding space, images as encoded matrices of numeric values indicating colors and shapes, and so on… Depending on the effectiveness, new variants of ICE combinations may be selected for the next iteration in various embodiments. The iterations may be continued until a termination criterion (such a determination that the marginal improvement in effectiveness metrics with new exploratory iterations has fallen below a threshold) is met in some embodiments. For some items, such as items for which the rates of consumer interactions fall below a threshold, or the iterations may be conducted indefinitely.”).
Zappella does not explicitly disclose […] using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance […]; and wherein the generative Al comprises a large language model or an image generative Al that generates at least one of text or an image based on the prompt.
However, Lasso further discloses:
[…] using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance […] (i.e. using machine learning model that uses regression analysis, gradient boosting, or neural networks to estimate/identify patterns or degrees of importance in the data) (Lasso: ¶ [0048] “The training service 321 can train a machine learning model or algorithm to identify anomalies. The machine learning algorithm or model can be any machine learning algorithm or model or combination thereof, including but not limited to nearest neighbor, support vector machines, gradient boosting, neural networks, logistic regression, linear regression, decision trees, random forest, Naive Bayes, k-means clustering, time series regression, pointwise prediction, stepwise regression, Gaussian mixture models, and means-shift clustering. The model can be any generative artificial learning model or algorithm, including but not limited to generative text models, such as conversational large language foundation models including ChatGPT ("generative pre-training transformer"), GPT-4, AlphaCode, GitHub Copilot, Bard, Cohere Generate, Claud, generative image and video models, such as Synthesia and DALL-E 2, and generative audio models.” Furthermore, as cited in ¶ [0063] “The weighting of the sub-scores can be determined by the machine learning algorithm discussed in reference to FIG. 5. The sub-scores can be weighted using a machine learning algorithm, including a classification machine learning algorithm. The anomaly impact score can be a rate represented by a number (e.g., 0 to 1) or a percentage (e.g., 0% to 100%). In some embodiments, the anomaly impact can be normalized to fit within a range of values (e.g., 0 to 1 or between 0 and 100).”); and
wherein the generative Al comprises a large language model or an image generative Al that generates at least one of text or an image based on the prompt (i.e. utilizing large language model or image generative AI to generate recommendations based on prompt or input data) (Lasso: ¶ [0045] “The anomaly service 318 can generate recommendations for proposed solutions and products based on the user profile data and the similarity data by using a machine learning model or algorithm. The machine learning algorithm or model can be any machine learning algorithm or model or combination thereof, including but not limited to nearest neighbor, support vector machines, gradient boosting, neural networks, logistic regression, linear regression, decision trees, random forest, Naive Bayes, k-means clustering, time series regression, pointwise prediction, stepwise regression, Gaussian mixture models, and means-shift clustering. The model can be any generative artificial learning model or algorithm, including but not limited to generative text models, such as conversational large language foundation models including GPT ("generative pre-training transformer"), including ChatGPT, GPT-4, AlphaCode, GitHub Copilot, Bard, Cohere Generate, Claud, generative image and video models, such as Synthesia and DALL-E 2, and generative audio models. The machine learning model or algorithm can receive solution and product data. The machine learning model or algorithm can use the solution and product data, the user profiles, and the similarity data to generate recommendations for proposed solutions and products. For example, the machine learning model or algorithm can make similar or identical recommendations to similar user profiles based on the similarity data.”).
Therefore, it would have been obvious to one of ordinary skill in the art, at the time the invention was made, to add Lasso’s using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance; and wherein the generative Al comprises a large language model or an image generative Al that generates at least one of text or an image based on the prompt to Zappella’s input the new prompt to the generative Al to cause the generative Al to generate the new generation information corresponding to the new prompt. One of ordinary skill in the art would have been motivated to do so because “A machine learning model or algorithm can be trained to identify anomalies. The model can be trained using previous network traffic messages, previously identified anomalies, and user feedback. The machine learning model can be applied to current network traffic messages to identify anomalies.” (Lasso: ¶ [0006]).
With respect to Claims 10 and 11:
All limitations as recited have been analyzed and rejected to claim 1. Claim 10 recites “An information processing method implemented by a computer, the information processing method comprising:” (Zappella: Col. 18 Lines 31-46) the steps performed by system claim 1. Claim 11 recites “A non-transitory computer readable storage medium having stored therein an information processing program that causes a computer to execute a process, the process comprising:” (Zappella: Cols. 18-19 Lines 61-19) the steps performed by system claim 1. Claims 10 and 11 do not teach or define any new limitations beyond claim 1. Therefore they are rejected under the same rationale.
With respect to Claim 2:
Zappella teaches:
The information processing apparatus according to claim 1, wherein the processor is configured to estimate a degree of importance of each of the plurality of items (i.e. estimates/calculates metrics of importances or prioritization of each items based on acquired information) (Zappella: Col. 6 Lines 36-62 “Categories of ICEs which may be presented in one or more contexts may include, among others, still images, videos, audio recordings, text collections, web links, and so on. In some embodiments in which several different presentation contexts are available for a given item, the different contexts may be prioritized relative to one another-e.g., if far more consumers purchased an item as a result of viewing a recommendation than as a result of viewing an item details page, the optimization of the recommendation interface may be prioritized higher than the optimization of the details page. Records of earlier interactions performed for similar items (or the same item) in the different contexts may be analyzed to prioritize among contexts in such embodiments. In at least one embodiment, the ICOS may provide an indication of the particular presentation context or contexts for which a plurality of ICEs should be provided by a content source---e.g., it may be the case that a wider variety of ICES are needed for exploration with respect to an item details page than are needed for a search response context. In various embodiments, different types of effectiveness or utility metrics may be generated for ICEs at the machine learning models used during the iterative optimization phase. Such metrics may include, for example, web link click count metrics, sales metrics, shopping cart insertion metrics, wish list insertion metrics, and/or session engagement length metrics.”).
With respect to Claim 3:
Zappella teaches:
The information processing apparatus according to claim 1, wherein the processor is configured to estimate a degree of importance of each combination of the plurality of items (i.e. estimates/calculates metrics of importances or prioritization of combination of items based on acquired information such as shopping cart or wish list combination of items) (Zappella: Col. 6 Lines 36-62 “Categories of ICEs which may be presented in one or more contexts may include, among others, still images, videos, audio recordings, text collections, web links, and so on. In some embodiments in which several different presentation contexts are available for a given item, the different contexts may be prioritized relative to one another-e.g., if far more consumers purchased an item as a result of viewing a recommendation than as a result of viewing an item details page, the optimization of the recommendation interface may be prioritized higher than the optimization of the details page. Records of earlier interactions performed for similar items (or the same item) in the different contexts may be analyzed to prioritize among contexts in such embodiments. In at least one embodiment, the ICOS may provide an indication of the particular presentation context or contexts for which a plurality of ICEs should be provided by a content source---e.g., it may be the case that a wider variety of ICES are needed for exploration with respect to an item details page than are needed for a search response context. In various embodiments, different types of effectiveness or utility metrics may be generated for ICEs at the machine learning models used during the iterative optimization phase. Such metrics may include, for example, web link click count metrics, sales metrics, shopping cart insertion metrics, wish list insertion metrics, and/or session engagement length metrics.”).
With respect to Claim 4:
Zappella teaches:
The information processing apparatus according to claim 1, wherein the processor is configured to estimate the degrees of importance of the plurality of items based on the real-time evaluation information that is information indicating evaluation on the service using a plurality of pieces of generation information each being generated by using the generative Al based on a corresponding prompt among a plurality of prompts each including a different combination of pieces of information on the plurality of items (i.e. acquiring effectiveness information on a service or items, wherein the effectiveness is determined using artificial intelligence / machine learning on a prompt that includes information of the items and wherein the prompts include different combinations of information on a plurality of items to determine effectiveness) (Zappella: Cols. 2-3 Lines 60-15 “In at least some embodiments, one or more of the techniques described herein may be employed at a largescale content provider such as an e-retailer which has an inventory comprising millions of items, where the producers of at least some of the items may have relatively limited expertise regarding the effectiveness of different types of content elements with respect to increasing sales or other metrics of interest, and where at least some types of informational content that can be provided for such items is somewhat subjective. Such item producers may be referred to as the "long tail" of the distribution of item producers in some embodiments. Although such item producers may have limited experience and/or resources with respect to promoting their own items, the e-retailing organization ( or some affiliated organization) may in some embodiments have access to substantial artificial intelligence or machine learning resources, as well as an extensive collection of data regarding the effectiveness of various types if information that has been presented to potential consumers in the past. As such, various types of experimentation and optimization techniques may be employed at thee-retailer to help at least some such less-experienced item producers achieve their business objectives.” Furthermore, as cited in Cols. 6-7 Lines 56-3 “In various embodiments, different types of effectiveness or utility metrics may be generated for ICEs at the machine learning models used during the iterative optimization phase. Such metrics may include, for example, web link click count metrics, sales metrics, shopping cart insertion metrics, wish list insertion metrics, and/or session engagement length metrics. In one embodiment, any of a number of granularities corresponding to respective target audiences of item consumers may be selected for optimizing the presentation of ICEs. The granularity levels may include, for example, global granularity (where all possible consumers are considered), group granularity (e.g., for potential consumers with some shared demographic or geographic characteristics), or individual granularity (for a single individual).”).
With respect to Claim 7:
Zappella teaches:
The information processing apparatus according to claim 1, wherein the processor is configured to: acquire the real-time evaluation information for each context of a user who uses the service (i.e. acquires information about interactions between the consumer and item) (Zappella: Col. 3 Lines 50-61 “In one embodiment, an analysis of the interactions of potential consumers of an item for which a particular set of one or more ICEs has been presented (or the lack of sufficient interactions with potential consumers) may help a service make a determination as to whether more ICE alternatives should be considered----e.g., a request to supply additional ICEs may be sent to a content source if a current set of ICEs fails to provide the desired business results. In some embodiments, a network-accessible service may enable item owners to query the service to determine whether their items should be considered candidates for ICE optimization.”),
estimate the degrees of importance of the plurality of items for each context of the user (i.e. estimates effectiveness or analyzes metrics associated with consumer in particular context or interaction) (Zappella: Col. 4 Lines 42-56 “In various embodiments, the particular optimization iteration may further comprise analyzing one or more metrics associated with the presentation of the particular set of ICEs (such as web-link clicks, purchases, and the like) to the audience in the particular context, and adjusting the ICE set if needed for the next iteration. In effect, in some embodiments, when the model is run in online mode, the utility or impact of a particular set of ICEs may be measured as soon as some interactions with potential consumers of the item are detected, and adjustments to the ICEs being presented (to the same audience, and/or to other audiences) may be made fairly rapidly based on the analysis of such interactions. Such a machine learning methodology may be referred to as an exploration-exploitation methodology in some embodiments.”), and
select information on an item that is to be included in the new prompt from among the pieces of information on the plurality of items for each context of the user (i.e. selects information to be included as prompt or input in machine learning model or AI based on consumer interactions with items) (Zappella: Col. 5 Lines 11-34 “A number of different techniques may be used to identify the baseline set in different embodiments---e.g., a machine learning model may be trained to predict an effective combination of ICE features using records of earlier consumer interactions with similar ICEs or similar items, a random collection of ICEs may be selected from among the available ICEs, and so on. After the baseline set has been identified, experiments may be carried out using variants of the baseline set in an ongoing manner, e.g., until some set of desired optimization objectives have been achieved using a machine learning model operating in an online or continuous learning mode, or until some optimization termination criteria are met. The termination criteria may, for example, include determining that the marginal improvements in effectiveness or utility being achieved by the variants has plateaued, that the overall consumption rate of the set of items for which ICE presentation is being optimized has fallen below some threshold which renders further optimizations inessential, and so on. In effect, the benefits of presenting various individual versions or variants of ICEs may be learned dynamically in an ongoing process, enabling the provider of the inventory to adjust to potentially changing trends in consumer behavior in at least some embodiments.”).
With respect to Claim 8:
Zappella teaches:
The information processing apparatus according to claim 1, wherein the evaluation on the service is at least one of evaluation that is made by a user who uses the service and evaluation that is based on a behavior of the user who uses the service (i.e. evaluates interactions between consumers and items, wherein the interactions reads on consumer behavior) (Zappella: Col. 3 Lines 50-61 “In one embodiment, an analysis of the interactions of potential consumers of an item for which a particular set of one or more ICEs has been presented (or the lack of sufficient interactions with potential consumers) may help a service make a determination as to whether more ICE alternatives should be considered----e.g., a request to supply additional ICEs may be sent to a content source if a current set of ICEs fails to provide the desired business results. In some embodiments, a network-accessible service may enable item owners to query the service to determine whether their items should be considered candidates for ICE optimization.” Furthermore, as cited in Col. 5 Lines 11-34 “A number of different techniques may be used to identify the baseline set in different embodiments---e.g., a machine learning model may be trained to predict an effective combination of ICE features using records of earlier consumer interactions with similar ICEs or similar items, a random collection of ICEs may be selected from among the available ICEs, and so on. After the baseline set has been identified, experiments may be carried out using variants of the baseline set in an ongoing manner, e.g., until some set of desired optimization objectives have been achieved using a machine learning model operating in an online or continuous learning mode, or until some optimization termination criteria are met. The termination criteria may, for example, include determining that the marginal improvements in effectiveness or utility being achieved by the variants has plateaued, that the overall consumption rate of the set of items for which ICE presentation is being optimized has fallen below some threshold which renders further optimizations inessential, and so on. In effect, the benefits of presenting various individual versions or variants of ICEs may be learned dynamically in an ongoing process, enabling the provider of the inventory to adjust to potentially changing trends in consumer behavior in at least some embodiments.”).
With respect to Claim 9:
Zappella teaches:
The information processing apparatus according to claim 1, wherein the service is an advertisement distribution service, and the generation information is an advertisement content that is distributed by the advertisement distribution service (i.e. service includes advertisements and optimization includes advertisement generation) (Zappella: Col. 13 Lines 49-55 “In one embodiment, ICE optimization may be performed for one or more forms of advertisements 722, such as advertisements sent via e-mails, social media tools, newspapers/periodicals, flyers and the like. In some embodiments, comparison tools 724, such as web pages which allow potential consumers to compare features of different items, may represent another presentation context for ICEs. ICE presentation optimization may be performed for one or more context not shown in FIG. 7 in some embodiments. In at least one embodiment, ICE presentation optimization may not necessarily be implemented for one or more of the contexts shown in FIG. 7.”).
Response to Arguments
Applicant’s arguments see pages 6-8 of the Remarks disclosed, filed on 03/12/2026, with respect to the 35 U.S.C. § 101 rejection(s) of claim(s) 1-4 and 7-11 have been considered but are not persuasive:
The Applicant asserts “However, the claims, as amended, now recite "using machine learning that uses regression analysis or gradient boosting to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables. " This specific machine learning implementation using regression analysis, gradient boosting, or neural networks with defined variable relationships cannot be practically performed in the human mind. Claims do not recite a mental process when they do not contain limitations that can practically be performed in the human mind, for instance when the human mind is not equipped to perform the claim limitations. See SRI Int 'l, Inc. V. Cisco Systems, Inc., 930 F.3d 1295, 1304 (Fed. Cir. 2019). Applicant notes that the claims further recite "wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt." This specifies that the generative AI is a large language model-which, as described in the specification, is "trained to estimate and output a next token from an input token string or an image generative AI that "generates an image from text. " Spec., p. 12, 11. 3-10, 19-23. These are specific technical implementations of generative AI that cannot be performed mentally or through pen and paper. The combination of regression analysis, gradient boosting, or neural networks with specific dependent and independent variable configurations, followed by inputting prompts to a large language model or image generative AI to generate new content, represents a specific technical implementation that is fundamentally different from abstract mental processes or mathematical concepts.” The Examiner respectfully disagrees. The amendments reciting –“ using machine learning that uses regression analysis or gradient boosting to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt” are seen as additional limitations to the abstract idea and are analyzed under step 2A prong II and/or step 2B. Furthermore, Independent Claims 1, 10, and 11 recite limitations directed to the abstract idea including “acquiring a prompt that includes pieces of information on a plurality of items; acquiring real-time evaluation information that is information indicating evaluation on a service using generation information; and estimating degrees of importance of the plurality of items based on the real-time evaluation information that is acquired; automatically generating a new prompt in real-time from one or more of the plurality of items based on the degrees of importance; and inputting the new prompt to generate the new generation information corresponding to the new prompt”. These further limitations are not seen as any more than the judicial exception. Claims 1, 10, and 11 recite additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. Generating a new prompt based on degrees of importance in order to generate new generation information is considered to be an abstract idea, specifically, certain methods of organizing human activity; such as managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules or instructions) because the claims are directed to managing a relationship between parties (i.e. evaluating a service) in order to estimate information (i.e. importance of items). Furthermore, generating a new prompt based on degrees of importance in order to generate new generation information is also considered to fall under another grouping of abstract idea, specifically, Mental Processes; such as concepts performed in the human mind (including an observation, evaluation, judgment, opinion) because the claims are directed to acquiring information (i.e. prompt and real-time evaluation of a service); estimating information (i.e. degrees of importance of items) based on the acquired information; generating a new prompt based on degrees of importance; and inputting the new prompt to generate new information corresponding to the new prompt which can all be performed in the human mind. Finally, generating a new prompt based on degrees of importance in order to generate new generation information is also considered to fall under another grouping of abstract idea, specifically, Mathematical Concepts; such as mathematical relationships, formulas or equations, and calculations, because the claims are directed to estimating degrees of importance of items based on an evaluation on a service which may require mathematical relationship in order to determine the estimation.
The Applicant also asserts “The Examiner has alleged that the additional limitations are merely "apply it " language or instructions to implement an abstract idea on a computer. Applicant respectfully disagrees. The amended limitation "using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables " as recited by claim 1 is not merely "applying " machine learning as a generic tool. Rather, it specifies particular machine learning techniques (regression analysis, gradient boosting, or neural networks) with a defined technical configuration (evaluation values as dependent variables and items as independent variables). Similarly, the amended limitation "wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt " as recited by claim 1 is not generic "apply it " language-it defines a specific type of generative AI with specific technical functionality. This directly responds to the Examiner's recommendation to "clarify how the generative AI generates a new prompt with technical details rooted in computer technology."”. The Examiner respectfully disagrees. The “machine learning” is recited in the claim limitations as – “estimating degrees of importance of the plurality of items based on the real-time evaluation information that is acquired at the acquiring using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables”. Here, the degrees of importance are estimated “using machine learning that uses regression analysis, gradient boosting, or neural networks” which clearly shows “apply it” language. “Generative AI” is recited in the claim limitations as – “acquiring real-time evaluation information that is information indicating evaluation on a service using generation information that is information that is generated by using generative AI based on the prompt; and inputting the new prompt to the generative AI to cause the generative AI to generate the new generation information corresponding to the new prompt, wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. Here, the generation information is generated “by using generative AI based on the prompt” and the prompt is inputted “to the generative AI to cause the generative AI to generate the new generation information corresponding to the new prompt” which clearly shows “apply it” language. The claims go further to describe the generative AI to be either a large language model or image generative AI to generate text or image, however, this further describes the computing environment in which the abstract idea takes place. Furthermore, Claims 1, 10, and 11 recite additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. The additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt” are not found to integrate the judicial exception into a practical application because they are seen as adding the words “apply it” (or an equivalent such as “using”) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f) and generally linking the use of the judicial exception to a particular technological environment or field of use – see MPEP 2106.05(h). Here, the “generative AI comprises a large language model or an image generative AI” is merely being used to describe the text or image or information that is generated and the “machine learning that uses regression analysis, gradient boosting, or neural networks” is merely being used to as a tool to estimate the degrees of importance of items. Accordingly, alone, and in combination, these additional elements are seen as using a computer or tool to perform an abstract idea, adding insignificant-extra-solution activity to the judicial exception. They do no more than link the judicial exception to a particular technological environment or field of use, i.e. generative AI or machine learning, and therefore do not integrate the abstract idea into a practical application. The courts decided that although the additional elements did limit the use of the abstract idea, the court explained that this type of limitation merely confines the use of the abstract idea to a particular technological environment and this fails to add an inventive concept to the claims (See Affinity Labs of Texas v. DirecTV, LLC,).
The Applicant finally asserts “The amended claims represent an integrated technical solution-a closed-loop optimization system for generative AI prompts. The system acquires real-time evaluation information, estimates degrees of importance using specific machine learning techniques (regression analysis, gradient boosting, or neural networks with defined variable relationships), automatically generates optimized prompts based on the estimated importance, and inputs those prompts to a specifically-defined generative Al (a large language model or image generative AI) to generate new content. This combination of elements provides a specific technical solution to the technical problem of optimizing prompts for generative AI systems, representing a practical application of any underlying concepts. Applicant's specific combination of regression analysis, gradient boosting, or neural networks with defined variable relationships, combined with a large language model or image generative AI, as recited by claims 1, 10, and 11 represents significantly more than a generic computer implementation. These are specific algorithmic techniques with defined data configurations operating in conjunction with specifically-defined generative AI systems. The ordered combination of acquiring real-time evaluation information, estimating degrees of importance using regression analysis, gradient boosting, or neural networks with specific variable configurations, automatically generating a new prompt based on the estimated importance, and inputting the new prompt to a large language model or image generative AI to cause generation of new text or images represents a specific technical process that amounts to significantly more than any alleged abstract idea.” The Examiner respectfully disagrees. The claims do not provide a solution to the technical problem of “optimizing prompts for generative AI systems, representing a practical application of any underlying concepts”. At most, the claims recite the limitation – “automatically generating a new prompt in real-time from one or more of the plurality of items based on the degrees of importance”, however, this generation of a new prompt is recited at such a high level (e.g. based on the estimated degrees of importance). Examiner recommends amending the claims to clarify how the new prompt is generated in detail and the how being rooted in computer technology. Furthermore, Claims 1, 10, and 11 recite additional limitations including “information that is generated by using generative AI based on the prompt; using machine learning that uses regression analysis, gradient boosting, or neural networks to estimate the degrees of importance while adopting evaluation values as dependent variables and items as independent variables; to the generative AI to cause the generative AI; and wherein the generative AI comprises a large language model or an image generative AI that generates at least one of text or an image based on the prompt”. These additional limitations do not integrate the judicial exception (abstract idea) into a practical application because of the analysis provided in Step 2A, Prong II. The independent claims do not include additional elements or a combination of elements that result in the claims amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional elements listed amount to no more than mere instructions to apply an exception using a generic computer component. In addition, the applicant’s specifications describe “general purpose graphic processing unit”, pages 28-31, for implementing the computer, which do not amount to significantly more than the abstract idea of itself, which is not enough to transform an abstract idea into eligible subject matter. There is no improvement in the functioning of the computer or technological field, and there is no transformation of subject matter into a different state. Therefore, the rejection(s) of claim(s) 1-4 and 7-11 under 35 U.S.C. § 101 is maintained above with an updated analysis.
Applicant’s arguments see pages 8-9 of the Remarks disclosed, filed on 03/12/2026, with respect to the 35 U.S.C. § 103 rejection(s) of claim(s) 1-4 and 7-11 over Zappella have been considered but are moot because the arguments do not apply to the new ground(s) of rejection is made in view of U.S. Publication 2025/0168056 to Lasso.
Conclusion
The prior art made of record and not relied upon is considered pertinent to Applicant’s disclosure. The following reference are cited to further show the state of the art:
U.S. Publication 2024/0427990 to Albasiri for disclosing a machine learning system to tokenize, classify, and generate representations from a provided input to provide a combined output. An input may be received and processed into tokens for classification based on semiotic classes. For a given semiotic class, particular rule-based algorithms may be selected to generate a desired output for a selected output language. An input may include an auditory or textual input, which may be in a different language from the selected output language, where the particular rule-based algorithms may include morphological rules for particular semiotic classes. Different rule-based algorithms may be modularly generated for particular languages and semiotic classes to build a library of models for processing different inputs.
U.S. Publication 2021/0073832 to Shido for disclosing a mechanism for facilitating appropriate pricing for an item to be listed for sale on an electronic commerce platform. In an information processing method, one or a plurality of processors included in an information processing device performs: obtaining first item information of a first item; performing sales evaluations for price ranges of the first item and obtaining evaluation information on the sales evaluations, the sales evaluations being performed based on sales information of second items matched with or similar to the first item, the second items being extracted by using the first item information from among items registered in an electronic commerce platform; and performing control to display the sales evaluations for the price ranges of the first item by using a plurality of regions identifiably on a screen, based on the evaluation information.
U.S. Publication 2021/0312495 to Hassan for disclosing a system and method for optimizing ad campaigns, which considers the relationship of items and immediately takes into account the future estimated impact of optimizations.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Azam Ansari, whose telephone number is (571) 272-7047. The examiner can normally be reached from Monday to Friday between 8 AM and 4:30 PM.
If any attempt to reach the examiner by telephone is unsuccessful, the examiner's supervisor, Waseem Ashraf, can be reached at (571) 270-3948.
Another resource that is available to applicants is the Patent Application Information Retrieval (PAIR). Information regarding the status of an application can be obtained from the (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAX. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pairdirect.uspto.gov. Should you have questions on access to the Private PAIR system, please feel free to contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
Applicants are invited to contact the Office to schedule either an in-person or a telephonic interview to discuss and resolve the issues set forth in this Office Action. Although an interview is not required, the Office believes that an interview can be of use to resolve any issues related to a patent application in an efficient and prompt manner.
/AZAM A ANSARI/
Primary Examiner, Art Unit 3621
June 10, 2026
Prosecution Insights