DETAILED ACTION
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 .
Status of the Claims
The Amendment filed on 12/19/2025 has been entered. Claims 1-20 are pending in the instant patent application. Claims 1-3, 5-8, 10, 14-16 and 18-20 are amended.
Response to Claim Amendments
Applicant’s amendments to the claims are insufficient to overcome the 35 U.S.C. §101 rejections. The rejections remain pending and are updated and addressed below in light of the amendments and per guidelines for 101 analysis (PEG 2019).
Applicant’s amendments to the claims are insufficient to overcome the 35 U.S.C. §103 rejections. The rejections remain pending and are updated and addressed below in light of the amendments and newly cited art.
Response to 35 U.S.C. §101 Arguments
Applicant’s arguments regarding 35 U.S.C. §101 rejection of the claims have been fully considered, but are not persuasive.
Regarding Applicant’s arguments that the claims are patent eligible under Step 2A Prong Two, Examiner respectfully disagrees. Examiner will note an important consideration to evaluate when determining whether the claim as a whole integrates a judicial exception into a practical application is whether the claimed invention improves the functioning of a computer or other technology. MPEP 2106.04(a) and 2106.05(a) provide a detailed explanation of how to perform this analysis. In short, first the specification should be evaluated to determine if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The specification need not explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. In analyzing the specification, Examiner maintains that the specification sets forth an improvement, but in a conclusory manner and furthermore the claims do not reflect the disclosed improvement or effectively demonstrate an improvement to existing technology. In addition, (ref: 2106.04(d)(1)). In addition, in light of the amendments, Examiner will state that the additional elements in the claim language individually as well as a whole, are generic in functionality, in their generic capacity and are merely being used as tools to carry out the abstract idea. Furthermore, the courts have found claims requiring a generic computer or nominally reciting a generic computer may still recite a mental process even though the claim limitations are not performed entirely in the human mind. It is for in at least these reasons, that the claim language is directed towards abstract ideas.
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.
Regarding Claims 1-13, they are directed to a method, however the claims are directed to a judicial exception without significantly more. Claims 1-13 are directed to the abstract idea of solution recommendations.
Performing the Step 2A Prong 1 analysis while referring specifically to independent Claim 1, claim 1 recites receiving a query indicating a request for a change that provides a solution; triggering a model to provide a list of one or more solutions that are responsive to the query; calculating a vector distance between one or more solutions provided by the model and solutions stored and the solutions are stored as first vector embeddings that are clustered based on similarity; validating the one or more solutions provided based on the calculated vector distance; preparing a list of solutions by applying a first model to the validated one or more solutions, the first model being trained to generate one or more scores based on customer data stored, and the customer data is stored as second vector embeddings that are clustered based on a region, a country, a company size, an industry type, and/or a sentiment value; and responding to the query with the recommended list of solutions.
These claim limitations fall within the Mental Processes grouping of abstract ideas for they are concepts that can be practically performed in the human mind (including an observation, evaluation, judgment, opinion). Furthermore, the mere recitation of a machine learning model does not take the claim out of Mental Processes and the courts have found claims requiring a generic computer or nominally reciting a generic computer may still recite a mental process even though the claim limitations are not performed entirely in the human mind.
Accordingly, the claim recites an abstract idea and dependent claims 2-13 further recite the abstract idea.
Regarding Step 2A Prong 2 analysis, the judicial exception is not integrated into a practical application. In particular the claim recites the elements of an existing system, a large language model, a product master database, a first machine learning model and a customer vector database. The existing system, large language model, a product master database, a first machine learning model and a customer vector database are merely generic computing devices and do not integrate the judicial exception into a practical application.
With respect to 2B, the claims do not include additional elements amounting to significantly more than the abstract idea. Claims 1, 5-6 and 13 include various elements that are not directed to the abstract idea under 2A. These elements include an existing system, a large language model, a product master database, at least one server, a first machine learning model, a customer vector database and the generic computing elements described in the Applicant's specification in at least Para 0045. These elements do not amount to more than the abstract idea because it is a generic computer performing generic functions. Furthermore, Claim 1 recites computer functions that the courts have recognized as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) (See MPEP 2106.05(d)(ii)…at least, Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).
Therefore, Claims 1, 5-6 and 13, alone or in combination, are not drawn to eligible subject matter as they are directed to abstract ideas without significantly more.
Regarding Claims 14-19, they are directed to a system, however the claims are directed to a judicial exception without significantly more. Claims 14-19 are directed to the abstract idea of solution recommendations.
Performing the Step 2A Prong 1 analysis while referring specifically to independent Claim 1, claim 1 recites receiving a query indicating a request for a change that provides a solution; triggering a model to provide a list of one or more solutions that are responsive to the query; calculating a vector distance between one or more solutions provided by the model and solutions stored and the solutions are stored as first vector embeddings that are clustered based on similarity; validating the one or more solutions provided based on the calculated vector distance; preparing a list of solutions by applying a first model to the validated one or more solutions, the first model being trained to generate one or more scores based on customer data stored, and the customer data is stored as second vector embeddings that are clustered based on a region, a country, a company size, an industry type, and/or a sentiment value; and responding to the query with the recommended list of solutions.
These claim limitations fall within the Mental Processes grouping of abstract ideas for they are concepts that can be practically performed in the human mind (including an observation, evaluation, judgment, opinion). Furthermore, the mere recitation of a machine learning model does not take the claim out of Mental Processes and the courts have found claims requiring a generic computer or nominally reciting a generic computer may still recite a mental process even though the claim limitations are not performed entirely in the human mind.
Accordingly, the claim recites an abstract idea and dependent claims 15-19 further recite the abstract idea.
Regarding Step 2A Prong 2 analysis, the judicial exception is not integrated into a practical application. In particular the claim recites the elements of at least one processor, at least one memory, an existing system, a large language model, a product master database, at least one server, a first machine learning model, a customer vector database. The at least one processor, at least one memory, an existing system, a large language model, a product master database, at least one server, a first machine learning model and a customer vector database are merely generic computing devices and do not integrate the judicial exception into a practical application.
With respect to 2B, the claims do not include additional elements amounting to significantly more than the abstract idea. Claims 14-15 and 18-19 includes various elements that are not directed to the abstract idea under 2A. These elements include at least one processor, at least one memory, an existing system, at least one processor, at least one memory, an existing system, a large language model, a product master database, at least one server, a first machine learning model, a customer vector database, a plurality of machine learning models and the generic computing elements described in the Applicant's specification in at least Para 0045. These elements do not amount to more than the abstract idea because it is a generic computer performing generic functions. Furthermore, Claim 14 recites computer functions that the courts have recognized as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) (See MPEP 2106.05(d)(ii)…at least, Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).
Therefore, Claims 14-15 and 18-19, alone or in combination, are not drawn to eligible subject matter as they are directed to abstract ideas without significantly more.
Regarding Claim 20, it is directed a non-transitory computer-readable storage medium, however the claim is directed to a judicial exception without significantly more. Claim 20 is directed to the abstract idea of solution recommendations.
Performing the Step 2A Prong 1 analysis while referring specifically to independent Claim 20, claim 20 recites receiving a query indicating a request for a change that provides a solution; triggering a model to provide a list of one or more solutions that are responsive to the query; calculating a vector distance between one or more solutions provided by the model and solutions stored and the solutions are stored as first vector embeddings that are clustered based on similarity; validating the one or more solutions provided based on the calculated vector distance; preparing a list of solutions by applying a first model to the validated one or more solutions, the first model being trained to generate one or more scores based on customer data stored, and the customer data is stored as second vector embeddings that are clustered based on a region, a country, a company size, an industry type, and/or a sentiment value; and responding to the query with the recommended list of solutions.
These claim limitations fall within the Mental Processes grouping of abstract ideas for they are concepts that can be practically performed in the human mind (including an observation, evaluation, judgment, opinion). Furthermore, the mere recitation of a machine learning model does not take the claim out of Mental Processes and the courts have found claims requiring a generic computer or nominally reciting a generic computer may still recite a mental process even though the claim limitations are not performed entirely in the human mind.
Accordingly, the claim recites an abstract idea.
Regarding Step 2A Prong 2 analysis, the judicial exception is not integrated into a practical application. In particular the claim recites the elements of an existing system, a large language model, a product master database, a first machine learning model and a customer vector database. The existing system, large language model, a product master database, a first machine learning model and a customer vector database are merely generic computing devices and do not integrate the judicial exception into a practical application.
With respect to 2B, the claims do not include additional elements amounting to significantly more than the abstract idea. Claim 20 includes various elements that are not directed to the abstract idea under 2A. These elements include an existing system, a large language model, a product master database, a first machine learning model, a customer vector database and the generic computing elements described in the Applicant's specification in at least Para 0045. These elements do not amount to more than the abstract idea because it is a generic computer performing generic functions. Furthermore, Claim 20 recites computer functions that the courts have recognized as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) (See MPEP 2106.05(d)(ii)…at least, Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).
Therefore, Claim 20 is not drawn to eligible subject matter as it is directed to abstract ideas without significantly more.
Response to 35 U.S.C. §103 Arguments
Applicant’s arguments regarding 35 U.S.C. §103 rejection of the claims have been fully considered, but are not persuasive. Furthermore, Applicant’s arguments are moot in light of the newly amended language.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1, 5-6, 14 and 18-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vigil (US 2023/0281570 A1) in view of Powar et al. (US 12,141,134 B1) in view of Alfke et al. (US 2025/0291782 A1) further in view of McCormick (US 2023/0105159 A1).
Regarding Claim 1, Vigil teaches the limitations of Claim 1 which state
receiving a query indicating a request for a change that provides a solution to an existing system (Vigil: Para 0061 via System 202 includes a query formation module 206 that helps the user generate context-aware queries when the user needs helps in managing or improving resources. Module 206 can assist the user from the context-aware search query based on collected data or reported data. The context-aware search query includes identification of components, bug code (if any), model number, and similar other information used to precisely identify the issue. In an embodiment, module 206 helps the user form a context-aware query or add context to a query submitted by the user automatically based on the collected data about the issue. The query formation module 206 can take voice, text, or image query from the user, identify the relevant keywords from the user query and add other context-aware data to form the context-aware query that can be submitted to a search engine. The query formation module 206 can also receive the image-based search request);
However, Vigil does not explicitly disclose the limitations of Claim 1 which state triggering a large language model to provide a list of one or more solutions that are responsive to the query.
Powar though, with the teachings of Vigil, teaches of
triggering a large language model to provide a list of one or more solutions that are responsive to the query (Powar: Col 20 lines 34-50 via At step 104, a user query received by a Large Language Model (LLM), is obtained, wherein the user query pertains to a request for execution of a given task. At step 106, whether the given task matches the corresponding descriptor of any playbook from amongst the set of playbooks, is determined. When it is determined that the given task matches a given descriptor of a given playbook from amongst the set of playbooks, subsequently, at step 108, the LLM selects one or more external tools, wherein the one or more external tools read the plain text descriptions of the given set of sub-tasks. At step 110, a software framework associated with the LLM tracks and executes the given set of sub-tasks in the given playbook, using the one or more external tools, in coordination with the LLM. At step 112, the LLM generates a query response including results of the execution of the given set of sub-tasks, and sends the query response to the user device).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil with the teachings of Powar in order to have triggering a large language model to provide a list of one or more solutions that are responsive to the query. The motivations behind this being to incorporate the teachings of utilizing large language models for task planning as taught by Powar. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention.
In addition, Vigil does not explicitly teach the limitations of Claim 1 which state calculating a vector distance between the one or more solutions provided by the LLM and solutions stored in a product master database, wherein the product master database is a first vector database, and the solutions are stored as first vector embeddings that are clustered based on similarity.
Alfke though, with the teachings of Vigil/Powar, teaches of
calculating a vector distance between the one or more solutions provided by the LLM and solutions stored in a product master database, wherein the product master database is a first vector database, and the solutions are stored as first vector embeddings that are clustered based on similarity (Alfke: Para 0005, 0019, 0027-0028, 0033, 0035 via the vector index is stored in a database table. Each record of the database table stores a vector and an identifier of the record associated with the vector…The system performs vector searches associated with a target vector by finding a set of partitions that either include the target vector or are near the target vector. The system may select partitions near the target vector by selecting partitions having centroids closest to the target vector. The system determines the distances between the target vector and the centroids of the partitions and selects the k nearest vectors within the vectors of the nearest partitions…the vector index generator 210 performs clustering of vectors to group the vectors into “buckets” for efficiently processing queries based on vectors such as nearest neighbor queries. According to an embodiment, the system performs clustering of vectors using inverted files. The clustering may be performed using a library such as open-source FAISS (Facebook AI Similarity Search) but is not limited to any specific implementation…training is performed by selecting a sample of vectors and performing clustering to determine an initial set of centroids that are used to assign the remaining vectors to buckets, each bucket corresponding to a centroid storing vectors that are within a threshold distance of the centroid…if a large number of vectors are available initially, the system performs sampling to determine a subset of vectors and perform clustering of the vectors to determine a set of clusters. The system uses the centroids of the clusters for determining which clusters the remaining vectors or any vectors received subsequently belong. For example, each vector is assigned to the cluster corresponding to the centroid that is nearest to the vector…The system stores the cluster information in a vector index. According to an embodiment, the system uses centroids determined by the clustering to define a set of buckets that all the vectors are assigned to. The system stores information describing the clusters in a vector index. According to an embodiment, the vector index stores information describing the set of buckets in an ordered key-value store. The key-value store may represent a database table managed by an embedded database management system (DBMS)).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil/Powar with the teachings of Alfke in order to have calculating a vector distance between the one or more solutions provided by the LLM and solutions stored in a product master database, wherein the product master database is a first vector database, and the solutions are stored as first vector embeddings that are clustered based on similarity. The motivations behind this being to incorporate the teachings of vector searching in databases utilizing vector distances. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention.
The combination of Vigil/Powar/Alfke further teaches the limitation of Claim 1 which states
validating the one or more solutions provided by the LLM based on the calculated vector distance (Alfke: Para 0050, 0051 via According to an embodiment, the query is a nearest neighbor query and the system identifies 440 the k nearest vectors based on their distance from a target vector specified by the query. According to an embodiment, the system creates an empty result list of (distance, docID) tuples. The system keeps the list ordered by distance. The system maintains a maximum capacity of the list to be k, the number of results requested by the query. The system may use data structures such as a heap or tree to optimize the list…The distance to the target vector may be computed using distance metrics such as Euclidean distance or cosine distance. The system inserts the resulting (distance, docID) tuple into the result list, maintaining its sort order. If the list grows beyond size k, the system removes the last element to keep the size of the list to within k while keeping the list sorted. As the system completes iterating through all the vectors of all the clusters identified in the steps 420, the result list stores the k nearest vectors to the target vector).
Furthermore, Vigil does not explicitly disclose the limitation of Claim 1 which states preparing a recommended list of solutions by applying a first machine learning model to the validated one or more solutions, the first machine learning model being trained to generate one or more scores based on customer data stored in a customer vector database, wherein the customer vector database is a second vector database, and the customer data is stored as second vector embeddings that are clustered based on a region, a country, a company size, an industry type, and/or sentiment value.
McCormick though, with the teachings of Vigil/Powar/Alfke, teaches of
preparing a recommended list of solutions by applying a first machine learning model to the validated one or more solutions, the first machine learning model being trained to generate one or more scores based on customer data stored in a customer vector database, wherein the customer vector database is a second vector database, and the customer data is stored as second vector embeddings that are clustered based on a region, a country, a company size, an industry type, and/or sentiment value (McCormick: Para 0013 wherein 0013 summarizes the prior art as a whole, 0074-0077 via FIG. 8 illustrates an example process for automatically generating an output plan recommendation using a machine learning model, such as the personalized health plan recommendation model 122. At 704, control begins by obtaining a user profile table and corresponding supplemental data. At 708, control pre-processes the supplemental data to create a standardized feature vector. At 712, control then obtains a feature list for relevant health plan option for the individual. For example, relevant health plan options may be determined by identifying health plans offered by an employer of the individual, health plans offered at a location of the individual, general health plans offered by insurance companies that the individual is considering, etc. At 716, for each health plan feature, control associates a feature rank value based on corresponding attribute preferences from the user's profile table. An example feature rank value table is illustrated in FIG. 9A, and discussed further below. At 720, control uses a weighting factor for each feature rank value, based on an attribute weight list (such as the attribute weight list illustrated in FIG. 7B). An example table of weighted rank values is illustrated in FIG. 9B, and a table of resulting weighted attributes is illustrated in FIG. 9C. Control proceeds to calculate a profile match score for each health plan option at 724. For example, FIG. 9C illustrates example profile match scores for each of Plan A, Plan B and Plan C. The profile match scores may include a sum of the weighted, ranked attribute values. At 728, control generates a model output plan recommendation according to the profile match scores. For example, control may select one or more health plan options having a highest score or scores, for recommendation to the individual. FIG. 9D illustrates an example recommend plan table that may be presented to the individual. In various implementations, control may transform a user interface based on the recommendation output, to display the recommendation output to an individual. The recommendation output may be stored in a database. FIG. 9A illustrates an example table of ranked health plan feature scores, where values for each health plan attribute are ranked according to the individual's assigned attribute preferences. In the example of FIG. 9A, Plan A and Plan B each receive a rank score of 2 for the Plan Type category, while Plan C receives a rank score of 0. This is because the individual's survey responses indicated that they required an individual coverage plan).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil/Powar/Alfke with the teachings of McCormick in order to have preparing a recommended list of solutions by applying a first machine learning model to the validated one or more solutions, the first machine learning model being trained to generate one or more scores based on customer data stored in a customer vector database, wherein the customer vector database is a second vector database, and the customer data is stored as second vector embeddings that are clustered based on a region, a country, a company size, an industry type, and/or sentiment value. The motivations behind this being to incorporate the teachings of using machine learning that generate individualized recommendation outputs. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention
The combination of Vigil/Powar/Alfke/McCormick further teaches the limitation of Claim 1 which states
responding to the query with the recommended list of solutions (Vigil: Para 0078 via A search query builder/suggestive query module 510 helps the user build a context-aware query or provide suggestive query based on user-entered search parameters and data collected from resources. Processing of different types of searches is described later with reference to FIG. 6 and FIG. 7. An ML-based solution recommendation module 512 recommends the best solution and content to the user using a machine learning (ML) model. Module 512 present the recommended solution and content to a user-based submitted context-aware query and the profile of the user. Module 512 considers the type of solutions that were helpful for other users having matching profiles for similar situations/issues).
Regarding Claim 5, the combination of Vigil/Powar/Alfke/McCormick teaches the limitations of Claim 5 which state
wherein the triggering comprises providing a prompt to the LLM to provide a list of the one or more solutions responsive to the solution of the query (Powar: Col 20 lines 34-50 via At step 104, a user query received by a Large Language Model (LLM), is obtained, wherein the user query pertains to a request for execution of a given task. At step 106, whether the given task matches the corresponding descriptor of any playbook from amongst the set of playbooks, is determined. When it is determined that the given task matches a given descriptor of a given playbook from amongst the set of playbooks, subsequently, at step 108, the LLM selects one or more external tools, wherein the one or more external tools read the plain text descriptions of the given set of sub-tasks. At step 110, a software framework associated with the LLM tracks and executes the given set of sub-tasks in the given playbook, using the one or more external tools, in coordination with the LLM. At step 112, the LLM generates a query response including results of the execution of the given set of sub-tasks, and sends the query response to the user device).
Regarding Claim 6, the combination of Vigil/Powar/Alfke/McCormick teaches the limitations of Claim 6 which state
receiving, from the LLM, the list of the one or more solutions that are responsive to the query (Powar: Col 20 lines 34-50 via At step 104, a user query received by a Large Language Model (LLM), is obtained, wherein the user query pertains to a request for execution of a given task. At step 106, whether the given task matches the corresponding descriptor of any playbook from amongst the set of playbooks, is determined. When it is determined that the given task matches a given descriptor of a given playbook from amongst the set of playbooks, subsequently, at step 108, the LLM selects one or more external tools, wherein the one or more external tools read the plain text descriptions of the given set of sub-tasks. At step 110, a software framework associated with the LLM tracks and executes the given set of sub-tasks in the given playbook, using the one or more external tools, in coordination with the LLM. At step 112, the LLM generates a query response including results of the execution of the given set of sub-tasks, and sends the query response to the user device).
Regarding Claims 14 and 18-19, they are analogous to Claims 1 and 5-6 respectively and are rejected for the same reasons. See also Vigil: Para 0090.
Regarding Claim 20, it is analogous to Claim 1 and is rejected for the same reasons. See also Vigil: Para 0097.
Claim(s) 2-3 and 15-16 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vigil (US 2023/0281570 A1) in view of Powar et al. (US 12,141,134 B1) in view of Alfke et al. (US 2025/0291782 A1) in view of McCormick (US 2023/0105159 A1) further in view of Walker et al. (US 2023/0267527 A1).
Regarding Claim 2, while the combination of Vigil/Powar/Alfke/McCormick teaches the limitations of Claim 1, it does not explicitly disclose the limitations of Claim 2 which state wherein the first machine learning model is one of a plurality of machine learning models, each of which being generated based on different training data and configured to apply a different scoring scheme to the validated one or more solutions.
Walker though, with the teachings of Vigil/Powar/Alfke/McCormick, teaches of
wherein the first machine learning model is one of a plurality of machine learning models, each of which being generated based on different training data and configured to apply a different scoring scheme to the validated one or more solutions (Walker: Para 0042-0043, 0060 via the machine learning model application 150 performs operations associated with training a plurality of models using the training data set 140 to generate a plurality of recommended models, applying the validation data set 142 to generate a plurality of predictions from the plurality of recommended models, and applying the test data set 144 to a selected recommended model. The machine learning model application 150 may include utilizing one or more supervised machine learning models including a random forest, k-nearest neighbors, a matrix factorization, and factorization machines, etc…the training data set 140 may include a first set of historic user profiles. The training data set 140 may be related to a plurality of historic user inputs associated with preferences of one or more services or items from an entity. The training data set 140 may also include a plurality of metrics from the entity for the one or more services or items…the computing device may execute operations for each of the one or more services or items to include selecting a recommended model from the plurality of recommended models based on the relevancy score of the selected metric or a combination of selected metrics. For example, as shown in FIG. 3, the computing device may include the score selection module 370x that selects a recommended model (e.g., Selected Model 384) from the plurality of recommended models 324 based on the relevancy score of the selected metric or a combination of selected metrics. For example, as shown in flow diagram 470 of FIG. 4B, the computing device may include the score selection module 370x that selects a recommended model from the plurality of recommended models 324 based on the relevancy score (e.g., Scored Data V 450v) of the selected metric (e.g., Precision 410v) or a combination of selected metrics (e.g., Precision 410v and Coverage 410x))).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil/Powar/Alfke/McCormick with the teachings of Walker in order to have wherein the first machine learning model is one of a plurality of machine learning models, each of which being generated based on different training data and configured to apply a different scoring scheme to the validated one or more solutions. The motivations behind this being to incorporate the teachings of automatically obtaining item-based recommendations that are most relevant through the automatic training and selection of multiple recommendation systems as taught by Walker. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention.
Regarding Claim 3, the combination of Vigil/Powar/Alfke/McCormick/Walker, teaches the limitations of Claim 3 which state
wherein the first machine learning model is selected as most accurate model among the plurality of machine learning models, and the first machine learning models is applied to the validated one or more solutions responsive to the selection (Walker: Para 0060 via For block 230 in flow diagram 200 of FIG. 2A and for block 250 in flow diagram 200 of FIG. 2B, the computing device may execute operations for each of the one or more services or items to include selecting a recommended model from the plurality of recommended models based on the relevancy score of the selected metric or a combination of selected metrics. For example, as shown in FIG. 3, the computing device may include the score selection module 370x that selects a recommended model (e.g., Selected Model 384) from the plurality of recommended models 324 based on the relevancy score of the selected metric or a combination of selected metrics. For example, as shown in flow diagram 470 of FIG. 4B, the computing device may include the score selection module 370x that selects a recommended model from the plurality of recommended models 324 based on the relevancy score (e.g., Scored Data V 450v) of the selected metric (e.g., Precision 410v) or a combination of selected metrics (e.g., Precision 410v and Coverage 410x))).
Regarding Claims 15-16, they are analogous to Claims 2-3 respectively and are rejected for the same reasons.
Claim(s) 4 and 17 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vigil (US 2023/0281570 A1) in view of Powar et al. (US 12,141,134 B1) in view of Alfke et al. (US 2025/0291782 A1) in view of McCormick (US 2023/0105159 A1) further in view of Prasanna (US 2011/0270646 A1).
Regarding Claim 4, while the combination of Vigil/Powar/Alfke/McCormick teaches the limitations of Claim 1, it does not explicitly disclose the limitations of Claim 4 which state wherein the query is received with the region, the country, the company size, and/or the industry type associated with the solution to the existing system.
Prasanna though, with the teachings of Vigil/Powar/Alfke/McCormick, teaches the limitations of Claim 4 which state
wherein the query is received with the region, the country, the company size, and/or the industry type associated with the solution to the existing system (Prasanna: Para 0977 via The output analyzer shown in FIG. 67 can not only display the output in a graphical form but the user can select parts of the solution in which he/she is interested and view only those. The user can zoom in or zoom out on any part of the solution. There is a query engine to help the user do this. The user can type in a query that works as a filter and shows only certain portions, satisfying the query (a query is a general Backus-Naur-Panini form specifiable expression composed of atomic operators). The module has the capability of clustering similar nodes and showing a simplified structure for better comprehension. The clustering can be done on many criteria such as geographic location, capacity etc. and can be chosen by the user. This makes a large, difficult to comprehend structure into a simplified easy to analyze structure).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil/Powar/Alfke/McCormick with the teachings of Prasanna in order to have wherein the query is received with the region, the country, the company size, and/or the industry type associated with the solution to the existing system. The motivations behind this being to incorporate the teachings of providing specific recommendations from the results. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention.
Regarding Claim 17, it is analogous to Claim 4 and is rejected for the same reasons.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vigil (US 2023/0281570 A1) in view of Powar et al. (US 12,141,134 B1) in view of Alfke et al. (US 2025/0291782 A1) in view of McCormick (US 2023/0105159 A1) further in view of Buhkin et al. (US 2014/0278264 A1).
Regarding Claim 7, while Vigil/Powar/Alfke/McCormick teaches the limitations of Claim 1, it does not explicitly disclose the limitation of Claim 7 which states wherein the validating comprises matching the list of the one or more solutions provided by the LLM with the solutions found in the product master database based on the calculated vector distance and eliminating from the list any of the one or more solutions that do not have a match in the product master database.
Buhkin though, with the teachings of Vigil/Powar/Alfke/McCormick, teach the limitations of Claim 7 which state
wherein the validating comprises matching the list of the one or more solutions provided by the LLM with the solutions found in the product master database based on the calculated vector distance and eliminating from the list any of the one or more solutions that do not have a match in the product master database (Buhkin: Para 0027 via Once filters are selected, the system moves on to block 225 where the filters are operated to choose filtered recommendations. In some embodiments the system receives an initial set of recommendations which may be all the recommendations available to the system, a predetermined regional set of recommendations, or a default starting recommendation set. In other embodiments, filters may generate query parameters for a database to select recommendations relevant to the user profile. The result is a filtered set of recommendations. In some embodiments, the system may apply further filters to the initial recommendation set to further remove, add, or modify recommendations, such as ordering the initial recommendation set to correspond to better matches being provided first).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil/Prasanna, with the teachings of Buhkin in order to have wherein the validating comprises matching the list of the one or more solutions provided by the LLM with the solutions found in the product master database based on the calculated vector distance and eliminating from the list any of the one or more solutions that do not have a match in the product master database. The motivations behind this being to incorporate the teachings of delivering personalized solutions to a user. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention.
Claim(s) 8-13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Vigil (US 2023/0281570 A1) in view of Powar et al. (US 12,141,134 B1) in view of Alfke et al. (US 2025/0291782 A1) in view of McCormick (US 2023/0105159 A1) further in view of Chun et al. (US 2014/0180999 A1).
Regarding Claim 8, while the combination of Vigil/Powar/Alfke/McCormick, teaches the limitations of Claim 1, it does not explicitly disclose the limitations of Claim 8 which state wherein the responding further comprises including the one or more scores for the recommended list of solutions.
Chun though, with the teachings of Vigil/Powar/Alfke/McCormick, teaches the limitation of Claim 8 which states
wherein the responding further comprises including the one or more scores for the recommended list of solutions (Chun: Para 0057 via Correspondingly, the help platform may receive the assessments on the various solutions from other users to determine assessment scores of the solutions. The assessment scores may be quantized values obtained by collecting and analyzing the assessments of the other users. Thus, when determining recommendation degrees of the solutions, the platform may take the assessment scores above as one of factors considered. In an embodiment, the recommendation degree of a solution may be determined as a weighted summation of occurrence frequency and assessment score, where the respective weights of the summations may be set based on implementation, which may be based on design choices. Thereby, the occurrence frequency of a solution and associated assessments of users may be considered in combination to give the recommendation degree of a solution, so that the recommendation degree may reflect the potential correctness of the solution).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Vigil/Powar/Alfke/McCormick with the teachings of Chun in order to have wherein the responding further comprises including the one or more scores for the recommended list of solutions. The motivations behind this being to incorporate the teachings of optimizing the degree of recommendations. Furthermore, in addition to being in the same CPC class, the teachings, suggestions, and motivations in this prior art would have led one of ordinary skill to modify the prior art reference or combine prior art reference teachings to arrive at the claimed invention.
Regarding Claim 9, the combination of Vigil/Powar/Alfke/McCormick/Chun, teaches the limitations of Claim 9 which state
wherein the one or more scores are each generated using a weighted scoring (Chun: Para 0057 via Correspondingly, the help platform may receive the assessments on the various solutions from other users to determine assessment scores of the solutions. The assessment scores may be quantized values obtained by collecting and analyzing the assessments of the other users. Thus, when determining recommendation degrees of the solutions, the platform may take the assessment scores above as one of factors considered. In an embodiment, the recommendation degree of a solution may be determined as a weighted summation of occurrence frequency and assessment score, where the respective weights of the summations may be set based on implementation, which may be based on design choices. Thereby, the occurrence frequency of a solution and associated assessments of users may be considered in combination to give the recommendation degree of a solution, so that the recommendation degree may reflect the potential correctness of the solution).
Regarding Claim 10, the combination of Vigil/Powar/Alfke/McCormick/Chun, teaches the limitations of Claim 10 which state
wherein the weighted scoring is based on a similarity score between each of the one or more solutions and the solutions included in the product master database (Chun: Para 0035 via when a questioning user faces a software problem, he can post the problem to a help platform, and the help platform may solicit or crowd source solutions for the software problem from other users. To help the user identify a correct solution from multiple solutions, semantic nodes may be introduced in the solutions submitted by other users to indicate meanings of user operations. Thus, information about similarity, correlation, or other characteristics. of solutions may be obtained by analyzing semantic nodes of the multiple solutions, so as to recommend a solution for the user based on the information).
Regarding Claim 11, the combination of Vigil/Powar/Alfke/McCormick/Chun, teaches the limitations of Claim 11 which state
wherein the weighted scoring is further based on a frequency of usage of the one or more solutions (Chun: Para 0057 via Correspondingly, the help platform may receive the assessments on the various solutions from other users to determine assessment scores of the solutions. The assessment scores may be quantized values obtained by collecting and analyzing the assessments of the other users. Thus, when determining recommendation degrees of the solutions, the platform may take the assessment scores above as one of factors considered. In an embodiment, the recommendation degree of a solution may be determined as a weighted summation of occurrence frequency and assessment score, where the respective weights of the summations may be set based on implementation, which may be based on design choices. Thereby, the occurrence frequency of a solution and associated assessments of users may be considered in combination to give the recommendation degree of a solution, so that the recommendation degree may reflect the potential correctness of the solution).
Regarding Claim 12, the combination of Vigil/Powar/Alfke/McCormick/Chun, teaches the limitations of Claim 12 which state
wherein the weighted scoring is further based on a customer sentiment indicating a rating provided by users of the one or more solutions (Chun: Para 0056 via Based on the initial recommendation degree determined from the occurrence frequency above, more factors may be taken into consideration to provide further recommendation degree information. In an embodiment, the method of this disclosure may further include obtaining assessments on the multiple solutions provided by other users. For example, as to the multiple solutions obtained at block 22, other users of the help platform may verify these solutions to determine whether the solutions can solve the software problem raised by the questioning user. After the verification, users may provide their assessments on the solutions to the help platform. The assessments may be embodied as, for example, ratings, scorings, votes etc. on the solutions).
Regarding Claim 13, the combination of Vigil/Powar/Alfke/McCormick/Chun, teaches the limitations of Claim 13 which state
wherein the customer sentiment is obtained using sentiment analysis obtained from at least one server or website (Chun: Para 0056 via Based on the initial recommendation degree determined from the occurrence frequency above, more factors may be taken into consideration to provide further recommendation degree information. In an embodiment, the method of this disclosure may further include obtaining assessments on the multiple solutions provided by other users. For example, as to the multiple solutions obtained at block 22, other users of the help platform may verify these solutions to determine whether the solutions can solve the software problem raised by the questioning user. After the verification, users may provide their assessments on the solutions to the help platform. The assessments may be embodied as, for example, ratings, scorings, votes etc. on the solutions).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Owen (US 2020/0279025 A1)
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to TYRONE E SINGLETARY whose telephone number is (571)272-1684. The examiner can normally be reached 9 - 5:30.
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, Beth Boswell can be reached at 571-272-6737. 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.
/T.E.S./Examiner, Art Unit 3625
/BETH V BOSWELL/Supervisory Patent Examiner, Art Unit 3625
Prosecution Insights