Detailed Action
This communication is in response to the Amendments and Arguments filed on 5/28/2026.
Claims 1,3-11 and 13-20 are pending and have been examined. Claims 2 and 12 have been cancelled.
Claims 1,3-11 and 13-20 are rejected. Hence this action has been made final.
Independent Claims 1, and 11 are System and Method claims, respectively.
Apparent priority: 10/16/2025.
Any previous objection/rejection not mentioned in this Office Action has been withdrawn by the Examiner.
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 .
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 5/28/2026, 1/22/2026 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Response to Amendment
The Applicants have amended the independent claims to include “intent, the data structure comprising a mapping between intent types and corresponding language models; receive, from a client device, a subsequent prompt corresponding to the intent relating to wager opportunities; elect, using the data structure that is updated based on the first score satisfying the assignment criterion, the first language model of the plurality of language models based on a classification of the subsequent prompt as corresponding to the intent; and generate, using the subsequent prompt and the first language model, output identifying at least one wager recommendation corresponding to the subsequent prompt.”.
Regarding the 35 USC § 101 rejection
Applicant notes the amended claim limitations limitations, taken as a whole, integrate any alleged abstract idea into a practical application by implementing a specific technical solution for dynamically selecting language models based on evaluated performance.
Applicant notes the specification describes concrete technical improvements achieved by the claimed technology. For example, the specification explains that "[a]s language models typically include a large number of parameters, invoking such language models typically requires significant processing resources that make language models challenging to use for certain applications (e.g., real-time or near real-time applications)." ( [0029]). The specification further explains that "[c]onventional systems that implement language models do not implement techniques for evaluating the performance of language models for different tasks or instructions, and therefore rely on sub-optimal, inefficient, and resource-constrained language models to perform different machine-learning tasks." ( [0035]). The claimed invention addresses these deficiencies by implementing "intent classification and input context generation to classify which language models are best suited to process a corresponding prompt without exhausting network or computing resources." ( [0035]).
Applicant notes The specification describes how these technical improvements are achieved. Specifically, the specification explains that "[b]y automatically selecting certain data to be included in the input context, the systems and methods described herein automatically limit the input context for the language model to a targeted subset of available data, thereby reducing the latency (e.g., processing time) and memory allocation required to carry out the requested operations using the language model." ( [0034]). The specification further explains that when "the evaluation score-9- of a language model satisfies or exceeds the assignment criterion, the data processing system can update the data structure that maps intents to language models" and "[o]nce the data structure is updated, the data processing system can select the appropriate language model for future input prompts based on the identified intent." ( [0128]-[0129]). These improvements "provide faster response times for multi-turn interactions, sustain throughput in high-load scenarios, and enable the use of large-scale language models within low-latency applications where conventional approaches would exceed performance constraints." ( [0032]).
Examiner notes the specification provides useful problem/solution narrative, but the eligibility inquiry is primarily directed to the claim language as a whole and whether the claim recites particular technical means that show an improvement in computer functionality or other technology
On the present claim wording, the limitations are largely functional and outcome oriented (evaluating model outputs, scoring language models, updating a mapping/data structure, classifying a prompt, and routing the prompt to a selected model) without concrete computational detail or a recitation of how the arrangements materially improve the functioning of the computer system itself (e.g., speed/latency reductions, memory or computational efficiency, novel data representations that reduce error by a measurable metric, or specific unconventional network architectures constrained in a way that produces the improvement.
Applicant notes the claims are not directed to an abstract idea, but rather to a specific technical solution that improves computer functionality by dynamically evaluating and selecting language models based on performance metrics and intent classification. Even assuming arguendo that the claims recite an abstract idea, the claims integrate that idea into a practical application by providing a technological improvement in how language models are selected and utilized in real-time computing environments.
Examiner notes the claim language is largely high level and functional these can be reasonably characterized as information processing/mental like steps (evaluating model outputs, scoring language models, updating a mapping/data structure, classifying a prompt, and routing the prompt to a selected model) Absent claim detail tying the operations to specific technical mechanisms that go beyond mere data processing, these limitations are susceptible to classification as mental/data manipulation concepts.
Examiner notes. Without claim specificity tying components to particular unconventional architectures, constrained parameterizations, training/regimen steps, or demonstrable improvements, the recited elements appear to be routine, conventional uses of language models and generic software components, and therefore fail to supply an inventive concept
Applicant arguments and amendments do not overcome the 35 U.S.C. § 101 rejection.
Applicant’s amendments and arguments with respect to the rejection(s) of claim(s) 1 and 11 under 35 U.S.C. § 103 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of Song (U.S. Patent Number US 20220188366 A1), and further in view of SHARMA (U.S. Patent Number US 20250238449 A1), and further in view of Jovanovic (U.S. Patent Number US 20250118156 A1).
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1,3-11 and 13-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
The independent Claims are directed to statutory categories:
Claim 1 is a System claim and directed to the machine or manufacture category of patentable subject matter.
Claim 7 is a Method claim and is directed to the machine or manufacture category of patentable subject matter.
Independent claim 1 recites,
“1. A system, comprising: one or more processors coupled to non-transitory memory, the one or more processors configured to:
maintain an evaluation dataset for a plurality of language models, the evaluation dataset comprising a first evaluation example including: (i) a respective input prompt indicating an intent relating to wager opportunities, and (ii) a respective output message identifying information associated with the intent and a corresponding wager recommendation; (This relates to a human pen and paper to maintain a dataset .)
generate, using the plurality of language models, a plurality of candidate outputs using a plurality of input prompts and the respective input prompt of the first input example; (This relates to a human using pen and paper to generate a plurality of outputs.)
determine a plurality of evaluation scores for the plurality of language models based on the respective output message of the first evaluation example and the plurality of candidate outputs, a first score of the plurality of scores corresponding to a first language model of the plurality of language models; (This relates to a human using logic and reasoning in the human mind to determine evaluation scores.)
and update, based on the first score satisfying an assignment criterion, a data structure to assign the first language model to input prompts identifying the intent, the data structure comprising a mapping between intent types and corresponding language models; (This relates to a human using pen and paper to update the data structure.)
receive, from a client device, a subsequent prompt corresponding to the intent relating to wager opportunities; (This relates to a human receiving a prompt using pen and pencil.)
select, using the data structure that is updated based on the first score satisfying the assignment criterion, the first language model of the plurality of language models based on a classification of the subsequent prompt as corresponding to the intent; and (This relates to a human selecting a language model.)
generate, using the subsequent prompt and the first language model, output identifying at least one wager recommendation corresponding to the subsequent prompt. (This relates to a human generating a wager recommendation corresponding to a prompt using pen and paper.)
The Dependent Claims do not include additional limitations that could incorporate the abstract idea into a practical application or cause the Claim as a whole to amount to significantly more than the underlying abstract idea.
Regarding Independent Claim 11, claim 1 is a method claim with limitations similar to that of Claim 1 and is rejected under the same rational.
This judicial exception is not integrated into a practical application. In particular, claims 1 and 11 recite additional elements of “processor” and “memory” For example, in [0165] of the as filed specification, there is description of using the elements of a computer include a processor for performing actions in accordance with instructions and one or more memory devices for storing instructions and data. Accordingly, these additional elements do not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. The claims are directed to an abstract idea.
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into a practical application, the additional element of using a processor and memory is noted as a general computer. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. Further, the additional limitation in the claims noted above are directed towards insignificant solution activity. The claims are not patent eligible.
Dependent claim 3 recites,
“3. The system of claim 1, wherein the one or more processors are further configured to: determine the plurality of evaluation scores based on a semantic similarity between the respective output message of the first evaluation example and the plurality of candidate outputs. (This relates to a human using logic and reasoning to determine scores.) No additional limitations present.
Dependent claim 4 recites,
“4. The system of claim 1, wherein the respective input prompt of the first evaluation example comprises a plurality of wager recommendations and the respective output message comprises a first wager recommendation of the plurality of wager recommendations.” (This relates to the type of input prompt received by a human and the output message using speech.) No additional limitations present.
Dependent claim 5 recites,
“5. The system of claim 4, wherein the one or more processors are further configured to: maintain a plurality of historical wager opportunities; (This relates to a human using pen and paper to maintain historical wager opportunities.)
and generate the plurality of wager recommendations based on a search operation of the plurality of wager opportunities using the respective input prompt.” (This relates to a human generating wager recommendations using pen and paper) No additional limitations present.
Dependent claim 7 recites,
“7. The system of claim 1, wherein the one or more processors are further configured to: determine a second plurality of evaluation scores for the plurality of language models using a second evaluation example corresponding to a second intent; (This relates to a human determining evaluation scores in the human mind using logic and reasoning) and assign a second language model of the plurality of language models to prompts corresponding to the second intent. (This relates to a human assigning a model using logic and reasoning in the human mind.) No additional limitations present.
Dependent claim 6 recites,
“6. The system of claim 4, wherein the one or more processors are further configured to: determine a first evaluation score of the plurality of evaluation scores corresponding to the first language model based on a comparison of the first wager recommendation and a corresponding wager recommendation included in a respective candidate output generated by the first language model. (This relates to a human using logic and reasoning in the human mind to determine socres) No additional limitations present.
Dependent claim 8 recites,
“8. The system of claim 1, wherein the intent identifies one or more of a wager type, a live event type, a team identifier, or an athlete identifier. (This relates to a human identifying intent using logic and reasoning.) No additional limitations present.
Dependent claim 9 recites,
“9. The system of claim 1, wherein the one or more processors are further configured to: generate a plurality of candidate outputs using a combination of the plurality of input prompts and the plurality of historical wager opportunities; (This relates to a human generating outputs using natural language understanding and logic and reasoning and pen and paper.)
and determine the plurality of evaluation scores based on the respective output message of a third evaluation example and the plurality of candidate outputs generated using the combination. (This relates to a human using logic and reasoning in the human mind to determine a score.) No additional limitations present.
Dependent claim 10 recites,
“10. The system of claim 1, wherein the one or more processors are further configured to: maintain a player profile associated with a client device; (This relates to a human using pen and paper to maintain a player profile.)
and generate the plurality of candidate outputs based on the player profile and the respective input prompt. (This relates to a human using pen and paper and logic and reasoning to generate candidate outputs.) No additional limitations present.
Regarding dependent Claim 13, claim 13 is a method claim with limitations similar to that of Claim 3 and is rejected under the same rational.
Regarding dependent Claim 14, claim 14 is a method claim with limitations similar to that of Claim 4 and is rejected under the same rational.
Regarding dependent Claim 15, claim 15 is a method claim with limitations similar to that of Claim 5 and is rejected under the same rational.
Regarding dependent Claim 16, claim 16 is a method claim with limitations similar to that of Claim 6 and is rejected under the same rational.
Regarding dependent Claim 17, claim 17 is a method claim with limitations similar to that of Claim 7 and is rejected under the same rational.
Regarding dependent Claim 18, claim 18 is a method claim with limitations similar to that of Claim 8 and is rejected under the same rational.
Regarding dependent Claim 19, claim 19 is a method claim with limitations similar to that of Claim 9 and is rejected under the same rational.
Regarding dependent Claim 20, claim 20 is a method claim with limitations similar to that of Claim 10 and is rejected under the same rational.
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 1,3-11 and 13-20 are rejected under 35 U.S.C. 103 as being unpatentable over Song (U.S. Patent Number US 20220188366 A1), and further in view of SHARMA (U.S. Patent Number US 20250238449 A1), and further in view of Jovanovic (U.S. Patent Number US 20250118156 A1).
Regarding independent Claim 1, Song teaches
1. A system, comprising: one or more processors coupled to non-transitory memory, the one or more processors configured to: (See Song [0011] “…The system also includes processor hardware configured to execute the instructions.”) maintain an evaluation dataset for a plurality of language models, (See Song [0083] “In various implementations, scaling from several thousand to up to one billion (or more) new articles may be implemented using search algorithms designed for fast searching capabilities. For example, a dataset may be segmented into pieces, with Voronoi cells defined in a d-dimensional space, and each database vector falling within one of the cells. During a search, only the database vectors y that are contained in the cell that the query x falls in, and a few neighboring ones, may be compared against the query vector.”) the evaluation dataset comprising a first evaluation example including: (i) a respective input prompt indicating an intent relating to wager opportunities, (See Song [0127] “Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”) and (ii) a respective output message identifying information associated with the intent and a corresponding wager recommendation; generate, using the plurality of language models, a plurality of candidate outputs using a plurality of input prompts and the respective input prompt of the first input example; (See Song [0009] “In other features, the method includes, in response to a determination that an elapsed time since generating the ranked element recommendation output satisfies recommendation time period criteria, obtaining updated current article database elements, obtaining an updated set of inputs specific to the individual entity, and processing the updated current article database elements and the updated set of inputs with the wide machine learning model and the deep machine learning model to generate an updated ranked element recommendation output.”) determine a plurality of evaluation scores for the plurality of language models based on the respective output message of the first evaluation example and the plurality of candidate outputs, a first score of the plurality of scores corresponding to a first language model of the plurality of language models; (see Song [0014] In other features, the instructions further include, in response to a determination that an elapsed time since generating the ranked element recommendation output is greater than a recommendation time period threshold, obtaining updated current article database elements, obtaining an updated set of inputs specific to the individual entity, and processing the updated current article database elements and the updated set of inputs with the wide machine learning model and the deep machine learning model to generate an updated ranked element recommendation output.”) (see Song [0015] In other features, the instructions further include at least one of determining a freshness score for a first one of the retrieved article database element according to a time period between a time when a recommendation API is called and a time that the retrieved article database element was published, determining an entity-access score for the retrieved article database element according to a time period since the retrieved article database element was accessed by the individual entity, determining an access score for the retrieved article database element according to number of overall accesses of the retrieved article database element by multiple entities, determining a section score according to an overlap between a section of the retrieved article database element and a section of at least one of the element access records associated with the individual entity, and determining a byline score according to an overlap between a byline of the retrieved article database element and a byline of at least one of the element access records associated with the individual entity. In other features, the instructions further include calculating an overall rank score for the retrieved article database element according to a combination of the freshness score, the entity-access score, the access score, the section score and the byline score.”)
Song does not specifically teach and update, based on the first score satisfying an assignment criterion, a data structure to assign the first language model to input prompts identifying the intent, the data structure comprising a mapping between intent types and corresponding language models; However, SHARMA does teach this limitation (see SHARMA [0118-0120] To illustrate, in some implementations, the causal query system 206 has access to multiple causal analysis models. In these implementations, the causal query system 206 may generate a causal script prompt (or a system prompt) that lists the models along with their capabilities. The causal query system 206 instructs the large generative model 230 to determine which of the causal analysis models is best suited to determine an outcome or answer for the causal question in the causal query. In some instances, selecting a causal analysis model may be based on a separate prompt and response from the large generative model 230. In other implementations, when only one causal analysis model is available, the causal query system 206 may select the default causal analysis model or skip selecting a model. [0119] With a causal analysis model selected, the causal query system 206 instructs the large generative model 230 in the causal script prompt to generate code for determining the effect of the causal question (e.g., determine the causal outcome based on the treatment included in the causal query). For example, the causal script prompt includes instructions to perform a causal analysis and/or prediction using the mapped causal graph or modified causal graph. [0120] As an example, for a causal analysis model that implements an object-oriented programming approach, the causal script prompt may include instructions to perform the causal analysis using one or more libraries and/or functions, create a model object from the provided data, identify a treatment and outcome variables, and estimate the causal effect using one or more approaches and/or counterfactual reasoning. In addition, the causal script prompt may include an example code for the large generative model 230 to follow in generating the script to provide to the causal analysis model.”) receive, from a client device, a subsequent prompt corresponding to the intent relating to wager opportunities; (see Song [0036] The system may use additional user attributes such … the device”) (see Song [0016] “In other features, merging the database elements includes, for each database elements of the wide ranked element list and the deep ranked element list, removing the item from the ranked element recommendation output in response to a determination that an item retrieval score of the database element is greater than a retrieval maximum threshold or lower than a retrieval minimum threshold, and removing the item from the ranked element recommendation output in response to a determination that an overall rank score of the database element is greater than a rank maximum threshold or less than a rank minimum threshold.”) generate, using the subsequent prompt and the first language model, output identifying at least one wager recommendation corresponding to the subsequent prompt. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
Song and SHARMA are in the same field of endeavor of signal processing, therefore, it would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of Song to incorporate and update, based on the first score satisfying an assignment criterion, a data structure to assign the first language model to input prompts identifying the intent of SHARMA. This allows for improved recommendations provided by the model as recognized by SHARMA [0035].
Song in view of SHARMA do not specifically teach select, using the data structure that is updated based on the first score satisfying the assignment criterion, the first language model of the plurality of language models based on a classification of the subsequent prompt as corresponding to the intent; and However, Jovanovic does teach this limitation. (see Jovanovic [0023] “In examples, the one or more machine learning models 116 may be trained to generate a natural language expression using a set of labeled data. For example, a labeled data set which matches gaming information to a correct natural language expression which documents the winning outcomes of a placed wager in a plain language, an explanations of the wager, etc., may be used to train the one or more machine learning model(s) 116. By training the one or more machine learning models 116 accordingly, the one or more machine learning models 116 may be leveraged to generate natural language expressions which detail the gaming selection, win conditions, etc. The training process updates the machine learning model(s) 116 to generate one or more specialized machine learning models operable to efficiently translate gaming data into natural language expressions. The machine learning model(s) 116 may be trained periodically as the curation engine receives new gaming data, based upon feedback and selections received from the devices in response to presenting the natural language expressions, etc. The periodic training of the models helps to optimize the curation engine 102 to generate natural language expressions efficiently in and ensure that the natural language expressions generated by the machine learning model(s) 116 are helpful to users. For example, by periodically training the models based upon feedback (e.g., in the form of direct feedback from users, indirect feedback, such as the acceptance or rejection of a bet, etc.) the machine learning model(s) 116 can be updated to produce natural language expressions that are helpful to end users. Furthermore, periodic updating can ensure that the one or more machine learning model(s)”)
Song in view of SHARMA and Jovanovic are in the same field of endeavor of signal processing, therefore, it would have been prima facie obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified the system of combination of Song and Sharma to incorporate selecting, using the data structure that is updated based on the first score satisfying the assignment criterion, the first language model of the plurality of language models based on a classification of the subsequent prompt as corresponding to the intent of Jovanovic. This allows for improved responses in a timely manner as recognized by Jovanovic [0027].
Regarding Independent Claim 11, claim 1 is a method claim with limitations similar to that of Claim 1 and is rejected under the same rational.
As to Claim 3 Song in view of SHARMA and further in view of Jovanovic teaches 3. The system of claim 1,
Furthermore, Song teaches wherein the one or more processors are further configured to: determine the plurality of evaluation scores based on a semantic similarity between the respective output message of the first evaluation example and the plurality of candidate outputs. (see Song [0122] “In various implementations, the wide and deep machine learning models may be used to combat self-imposed filter bubbles, where articles are recommended outside of the content that the user normally views. For example, by quantifying semantics to find articles with similar content to what the user typically reads, it is also possible to recommend articles with “opposing” content. The user may configure how they would like the opposing content articles to be presented, which may include factual and/or opinion articles. In various implementations, a Word2Vec model may be used to identify terms that are opposing to text of articles that the user normally reads.”) (see Song [0126] “Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all.”)
As to Claim 4 Song in view of SHARMA and further in view of Jovanovic teaches 4. The system of claim 1,
Furthermore, Song teaches wherein the respective input prompt of the first evaluation example comprises a plurality of wager recommendations and the respective output message comprises a first wager recommendation of the plurality of wager recommendations. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
As to Claim 5 Song in view of SHARMA and further in view of Jovanovic teaches 5. The system of claim 4,
Furthermore, Song teaches wherein the one or more processors are further configured to: maintain a plurality of historical wager opportunities; and generate the plurality of wager recommendations based on a search operation of the plurality of wager opportunities using the respective input prompt. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
As to Claim 6 Song in view of SHARMA and further in view of Jovanovic teaches 6. The system of claim 4,
Furthermore, Song teaches wherein the one or more processors are further configured to: determine a first evaluation score of the plurality of evaluation scores corresponding to the first language model based on a comparison of the first wager recommendation and a corresponding wager recommendation included in a respective candidate output generated by the first language model. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
As to Claim 7 Song in view of SHARMA and further in view of Jovanovic teaches 7. The system of claim 1,
Furthermore, Song teaches wherein the one or more processors are further configured to: determine a second plurality of evaluation scores for the plurality of language models using a second evaluation example corresponding to a second intent; and assign a second language model of the plurality of language models to prompts corresponding to the second intent. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
As to Claim 8 Song in view of SHARMA and further in view of Jovanovic teaches 8. The system of claim 1,
Furthermore, Song teaches 8. The system of claim 1, wherein the intent identifies one or more of a wager type, a live event type, a team identifier, or an athlete identifier. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
As to Claim 9 Song in view of SHARMA and further in view of Jovanovic teaches 9. The system of claim 1,
Furthermore, Song teaches 9. The system of claim 1, wherein the one or more processors are further configured to: generate a plurality of candidate outputs using a combination of the plurality of input prompts and the plurality of historical wager opportunities; and determine the plurality of evaluation scores based on the respective output message of a third evaluation example and the plurality of candidate outputs generated using the combination. (see Song [0126-0127] “[0126] Additional contexts that may be suitable for using the wide and deep machine learning models may include movie recommendations, book recommendations, product recommendations, gambling opportunities, etc. For example, the models may receive inputs of available movies or books that the user may have the option to view or read, and then retrieve and rank items that have a best match to movies or books from the user's viewing or read history. The sliding window for movie or book recommendations may be longer than news items, or in some cases a sliding window may not be necessary at all. [0127] Similarly, the models may receive inputs of available products that the user may purchase, and then retrieve and rank items that have a best match to products from the user's purchase history. As another example, the models may receive inputs of available betting options (for example, current sporting event betting opportunities) that the user may participate in, and then retrieve and rank items that have a best match to events from the user's betting history.”)
As to Claim 10 Song in view of SHARMA and further in view of Jovanovic teaches 10. The system of claim 1,
Furthermore, Song teaches 10. The system of claim 1, wherein the one or more processors are further configured to: maintain a player profile associated with a client device; and generate the plurality of candidate outputs based on the player profile and the respective input prompt. (see Song [0093-0094] “At 508, control obtains user attributes and a read articles history for the user (sometimes referred to as an individual entity). Control then proceeds to 512 to determine whether the number of articles read is greater than a specified read count threshold. The specified read count threshold may be indicative that there are enough read article data points (for example, at least ten articles, at least fifty articles, etc.), to make an article recommendation based on the user's read history. [0094] If control determines at 512 that the number of articles read is not greater than the specified threshold, control obtains a popular item list based on recent overall click counts at 516. For example, if the user has not yet read enough articles to base recommendations off of a read history of the user, control may start by selecting articles that are most popular to the overall readership of the news organization. At 520, control sets the obtained poplar article item list as the model input.”)
Regarding dependent Claim 13, claim 13 is a method claim with limitations similar to that of Claim 3 and is rejected under the same rational.
Regarding dependent Claim 14, claim 14 is a method claim with limitations similar to that of Claim 4 and is rejected under the same rational.
Regarding dependent Claim 15, claim 15 is a method claim with limitations similar to that of Claim 5 and is rejected under the same rational.
Regarding dependent Claim 16, claim 16 is a method claim with limitations similar to that of Claim 6 and is rejected under the same rational.
Regarding dependent Claim 17, claim 17 is a method claim with limitations similar to that of Claim 7 and is rejected under the same rational.
Regarding dependent Claim 18, claim 18 is a method claim with limitations similar to that of Claim 8 and is rejected under the same rational.
Regarding dependent Claim 19, claim 19 is a method claim with limitations similar to that of Claim 9 and is rejected under the same rational.
Regarding dependent Claim 20, claim 20 is a method claim with limitations similar to that of Claim 10 and is rejected under the same rational.
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 KRISTEN MICHELLE MASTERS whose telephone number is (703)756-1274. The examiner can normally be reached M-F 8:30 AM - 5:00 PM.
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, Pierre Louis Desir can be reached at 571-272-7799. 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.
/KRISTEN MICHELLE MASTERS/Examiner, Art Unit 2659
/PIERRE LOUIS DESIR/Supervisory Patent Examiner, Art Unit 2659
Prosecution Insights