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 .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 12/19/2024, 11/25/2025, 03/26/2025, 06/01/2026 was filed. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
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.
Claim 20 is rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claim(s) does/do not fall within at least one of the four categories of patent eligible subject matter.
Claim 20 recites the limitation of “computer-readable media” and is not limited to non-transitory computer readable medium. The specification defines computer-readable media as including communication media, such as a signal or carrier wave. See paragraph [0160]. Accordingly, under the broadest reasonable interpretation, claim 20 encompasses transitory signals and carrier waves, which are non-statutory subject matter.
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claim(s) 1-13, 15-20 are rejected under 35 U.S.C. 103 as being unpatentable over GRINBERG (US 20250063083 A1, Filed Date: Dec. 29, 2023) in view of Zheng (US 20220415320 A1, Filed Date: May, 16, 2022).
Regarding independent claim 1, GRINBERG teaches: A system for managing agents providing feedback to electronic documents, the system comprising: (GRINBERG – [0143] Process 800 may represent a non-transitory computer-readable medium, a method, or a system for selection operations for a plurality of distinct AI agents.)
processing circuitry; (GRINBERG – [0041] Certain embodiments disclosed herein may also include a computing device for generating features for work collaborative systems, the computing device may include processing circuitry communicatively connected to a network interface and to a memory, wherein the memory contains instructions that, when executed by the processing circuitry,)
and computer readable media comprising instructions that, when executed, cause the processing circuitry to: (GRINBERG – [0143] FIG. 8 is a flowchart of an exemplary process (800) for selection operations for a plurality of distinct AI agents, that may be executed by at least one processor. Process 800 is discussed herein for explanatory purposes and is not intended to be limiting. Process 800 may represent a non-transitory computer-readable medium, a method, or a system for selection operations for a plurality of distinct AI agents.)
based on a determination that each change of a first plurality of changes to a first plurality of historical documents is associated with a first user (GRINBERG – [0059] Examples of metadata may include user profiles, roles and permissions, activity logs, usage indications, preferences and settings, user associations/relationships, user history or a combination thereof. [0103] Alternatively, the AI agents may be continually updated, modified, changed, or replaced, either by an existing AI agent, or by a developer. Selectable AI agents may include, for example, formula generators, planning generators, text generators, chat responders, and application finders. [0127] In this example, the predetermined AI agent may assess features such as the writing complexity, language, quality, and user history, when comparing the responses. [0143] [0137-0139] agent selection and updating based on user needs)
and that the first user is assigned a first persona, (GRINBERG – [0008-0009] The operations may further include comparing information associated with each of the received responses, and selecting at least one AI agent from the plurality of distinct AI agents based on the comparison. [0009] selecting a particular AI agent from a pool of a plurality of AI agents, Selecting appropriate AI agent by assessing user history)
generate a first agent for the first persona based on both the first plurality of changes and first source data; (GRINBERG – [0018] FIG. 8 is a flow chart of an exemplary process for selection operations for a plurality of distinct AI agents consistent with some embodiments of the present disclosure; [0103-0104] Some disclosed embodiments involve selecting an AI agent from a plurality of available AI agents by analyzing application code, context, developer information, user information, or a combination thereof.)
determine a first plurality of vector embeddings for the first agent based on the first source data; (GRINBERG − [0047] In some embodiments, machine learning algorithms (also referred to as machine learning models or artificial intelligence in the present disclosure) may be trained using training examples, for example in the cases described below. mathematical embedding algorithms, natural language processing algorithms, support vector machines, random forests, nearest neighbors algorithms,)
determine a second plurality of vector embeddings based on second source data; (GRINBERG − [0047] In some embodiments, machine learning algorithms (also referred to as machine learning models or artificial intelligence in the present disclosure) may be trained using training examples, for example in the cases described below. mathematical embedding algorithms, natural language processing algorithms, support vector machines, random forests, nearest neighbors algorithms,)
determine that the first source data is associated with a second persona different from the first persona; (GRINBERG – [0008] The operations may further include comparing information associated with each of the received responses, and selecting at least one AI agent from the plurality of distinct AI agents based on the comparison. [0126-0128] Comparing information refers to examining two or more sets of data, facts, or details to identify similarities, differences, patterns, and/or relationships between them. As described herein, comparing information may be done by analyzing the responses received from a plurality of AI agents to assess their utility or applicability for addressing a query. A comparison could be made by analyzing segments or portions of information associated with each of the received responses, or by analyzing the entirety of the data or information associated with one or more queries and a plurality of AI agents.)
and generate a second agent for the second persona based on the first source data. (GRINBERG – [0008] The operations may further include comparing information associated with each of the received responses, and selecting at least one AI agent from the plurality of distinct AI agents based on the comparison. [0126-0128] Comparing information refers to examining two or more sets of data, facts, or details to identify similarities, differences, patterns, and/or relationships between them. As described herein, comparing information may be done by analyzing the responses received from a plurality of AI agents to assess their utility or applicability for addressing a query. A comparison could be made by analyzing segments or portions of information associated with each of the received responses, or by analyzing the entirety of the data or information associated with one or more queries and a plurality of AI agents. as an example, in response to a query, one AI agent may send a response “available,” and another AI agent may send a response “not available.” In another example, one AI agent may send a response, in which a portion of the response is “applicable,” while another AI agent sends a response, with a portion that says “not applicable.”)
GRINBERG does not explicitly teach: determine a similarity value based on the first plurality of vector embeddings
However, Zheng teaches: determine a first plurality of vector embeddings for the first agent based on the first source data; determine a second plurality of vector embeddings based on second source data; (Zheng − [0014] FIG. 6 illustrates an example view of an embedding space. [0092] [0163-0165] FIG. 6 illustrates an example view of a vector space 600. In particular embodiments, an object or an n-gram may be represented in a d-dimensional vector space, where d denotes any suitable number of dimensions. Although the vector space 600 is illustrated as a three-dimensional space, this is for illustrative purposes only, as the vector space 600 may be of any suitable dimension.)
determine a similarity value based on the first plurality of vector embeddings and the second plurality of vector embeddings; based on the similarity value, (Zheng − [0165] In particular embodiments, the social-networking system 160 may calculate a similarity metric of vectors in vector space 600. A similarity metric may be a cosine similarity, a Minkowski distance, a Mahalanobis distance, a Jaccard similarity coefficient, or any suitable similarity metric. As an example and not by way of limitation, a similarity metric of {right arrow over (v.sub.1)} and {right arrow over (v.sub.2)} may be a cosine similarity)
Accordingly, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention to have modify GRINBERG with the vector embedding similarity techniques taught by Zheng because Zheng teaches a known method for comparing vector embeddings and determining similarity between information represented in an embedding space. Such modification would improve the accuracy and consistency of GRINBERG’s comparison process when selecting an appropriate AI agent.
Regarding dependent claim 2, depends on claim 1, GRINBERG teaches: generate, based on the second agent and content of an electronic document, feedback information for the second persona; (GRINBERG – [0110] The AI agent may be configured to accept user feedback to improve performance. The AI agent may allow for additional operations, such as “summarize data,” in which the document may be analyzed and its contents summarized. Additional options from the AI agent may include “complete text,” in which the AI agent may complete a body of text that was started or initiated by a user input.)
generate data for a user interface configured to display an indication of the feedback information for the second persona; (GRINBERG – [0141-0143] Consistent with some embodiments, the feedback signal is received from a user via a user interface. [0143] Examples of allowing a user to provide feedback on a user interface are shown as thumbs up and thumbs down icons in FIG. 6 and FIG. 7.)
and output, to a user device, the data for the user interface to cause the user device to display the user interface configured to display the indication of the feedback information for the second persona. (GRINBERG – [0140] Some disclosed embodiments involve receiving a feedback signal on the outputted response, wherein the feedback signal is either positive or negative; if the feedback signal is negative, selecting at least one other AI agent from the plurality of distinct AI agents; and outputting the response of the at least one other selected AI agent. A feedback signal in this context refers to a reaction to an outputted response. The reaction may be a response or information that is sent back to a system, process, or device associated with the output. The feedback signal is received when it is captured or obtained. In some embodiments, the signal may be used as a means for improvement, analysis, or decision making. [0143] Examples of allowing a user to provide feedback on a user interface are shown as thumbs up and thumbs down icons in FIG. 6 and FIG. 7.)
Regarding dependent claim 3, depends on claim 1, GRINBERG teaches: wherein the instructions further cause the processing circuitry to: generate, based on the second agent and content of an electronic document, feedback information for the second persona; and output an indication of the feedback information to a user device associated with the electronic document. (GRINBERG – [0140-0143] Some disclosed embodiments involve receiving a feedback signal on the outputted response, wherein the feedback signal is either positive or negative; if the feedback signal is negative, selecting at least one other AI agent from the plurality of distinct AI agents; and outputting the response of the at least one other selected AI agent. A feedback signal in this context refers to a reaction to an outputted response. The reaction may be a response or information that is sent back to a system, process, or device associated with the output. The feedback signal is received when it is captured or obtained. In some embodiments, the signal may be used as a means for improvement, analysis, or decision making.)
Regarding dependent claim 4, depends on claim 1, GRINBERG teaches: wherein the instructions further cause the processing circuitry to: generate an encoder based on external data associated with the first agent, wherein to determine the first plurality of vector embeddings for the first agent, the instructions cause the processing circuitry to generate, with the encoder, the first plurality of vector embeddings based on the first source data, and wherein to determine the second plurality of vector embeddings for the second agent, the instructions cause the processing circuitry to generate, with the encoder, the second plurality of vector embeddings based on the second source data. (GRINBERG − [0047] In some embodiments, machine learning algorithms (also referred to as machine learning models or artificial intelligence in the present disclosure) may be trained using training examples, for example in the cases described below. mathematical embedding algorithms, natural language processing algorithms, support vector machines, random forests, nearest neighbors algorithms, [0162] Classify, as used herein, refers to a natural language processing technique that involves labeling or organizing data or elements of a language. Inferring an intent refers to implicit or explicit intent identification to classify elements of language based on what the user wants to achieve, or what the user is perceived to want. Intent extraction, in which an encoder-decoder topology may be commonly used could be a method to infer an intent, for example. Analyzing details associated with the user may refer to personalizing the analysis based on characteristics of a user or user needs. For example, a user may request an AI agent to build an element in the platform for a specific purpose and the AI agent can analyze the input in order to understand the desired element the client wishes it to build and the set of functionalities the AI agent should have access to in order to fulfill the intention behind the input, albeit being absent from the input.)
Regarding dependent claim 5, depends on claim 4, GRINBERG teaches: wherein to generate the encoder, the instructions cause the processing circuitry to: determine vocabulary information based on the external data; and generate the encoder based on the vocabulary information. (GRINBERG − [0061] It involves the ability of AI systems to perform tasks that usually require human intelligence. Examples of those tasks may include comprehension of natural language, recognition of patterns or structures in a data set, or the generation of predictions/forecasts. AI functionality incorporates different techniques and technologies such as Natural Language Processing (NLP), Natural Language Generation (NLG), Machine Learning (ML), Neural Network (NN), Deep Learning (DL), Large Language Model (LLM), or Computer Vision. AI functionalities empower applications to analyze, interpret, and generate data, automate processes, optimize performance, and offer intelligent solutions for intricate problems.)
Zheng teaches: wherein to generate the encoder, the instructions cause the processing circuitry to: determine vocabulary information based on the external data; and generate the encoder based on the vocabulary information. (Zheng − [0120] In particular embodiments, the context engine 220 may help the entity resolution module 212 improve entity resolution. The entity resolution module 212 may additionally extract features from contextual information, which is accessed from dialog history between a user and the assistant system 140. The entity resolution module 212 may further conduct global word embedding, domain-specific embedding, and/or dynamic embedding based on the contextual information. The processing result may be annotated with entities by an entity tagger. Based on the annotations, the entity resolution module 212 may generate dictionaries. In particular embodiments, the dictionaries may comprise global dictionary features which can be updated dynamically offline.)
Accordingly, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention to have modify GRINBERG with the vector embedding similarity techniques taught by Zheng because Zheng teaches a known method for comparing vector embeddings and determining similarity between information represented in an embedding space. Such modification would improve the accuracy and consistency of GRINBERG’s comparison process when selecting an appropriate AI agent.
Regarding dependent claim 6, depends on claim 4, GRINBERG teaches: wherein the instructions further cause the processing circuitry to: determine whether to initiate the generation of the encoder based on the external data. (GRINBERG − [0047] In some embodiments, machine learning algorithms (also referred to as machine learning models or artificial intelligence in the present disclosure) may be trained using training examples, for example in the cases described below. mathematical embedding algorithms, natural language processing algorithms, support vector machines, random forests, nearest neighbors algorithms, [0162] Classify, as used herein, refers to a natural language processing technique that involves labeling or organizing data or elements of a language. Inferring an intent refers to implicit or explicit intent identification to classify elements of language based on what the user wants to achieve, or what the user is perceived to want. Intent extraction, in which an encoder-decoder topology may be commonly used could be a method to infer an intent, for example. Analyzing details associated with the user may refer to personalizing the analysis based on characteristics of a user or user needs. For example, a user may request an AI agent to build an element in the platform for a specific purpose and the AI agent can analyze the input in order to understand the desired element the client wishes it to build and the set of functionalities the AI agent should have access to in order to fulfill the intention behind the input, albeit being absent from the input.)
Regarding dependent claim 7, depends on claim 1, GRINBERG does not explicitly teach: wherein to determine the similarity value, the instructions cause the processing circuitry to: determine a cosine similarity value between the first plurality of vector embeddings and the second plurality of vector embeddings; or determine a Euclidean distance value between the first plurality of vector embeddings and the second plurality of vector embeddings
However, Zheng teaches: wherein to determine the similarity value, the instructions cause the processing circuitry to: determine a cosine similarity value between the first plurality of vector embeddings and the second plurality of vector embeddings; or determine a Euclidean distance value between the first plurality of vector embeddings and the second plurality of vector embeddings. (Zheng − [0165] In particular embodiments, the social-networking system 160 may calculate a similarity metric of vectors in vector space 600. A similarity metric may be a cosine similarity, a Minkowski distance, a Mahalanobis distance, a Jaccard similarity coefficient, or any suitable similarity metric. As an example and not by way of limitation, a similarity metric of {right arrow over (v.sub.1)} and {right arrow over (v.sub.2)} may be a cosine similarity)
Accordingly, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention to have modify GRINBERG with the vector embedding similarity techniques taught by Zheng because Zheng teaches a known method for comparing vector embeddings and determining similarity between information represented in an embedding space. Such modification would improve the accuracy and consistency of GRINBERG’s comparison process when selecting an appropriate AI agent.
Regarding dependent claim 8, depends on claim 1, GRINBERG teaches: wherein the second plurality of vector embeddings are for a third agent associated with a third persona different from the first agent; wherein to determine that the first source data is associated with the second persona, the instructions cause the processing circuitry to determine that the similarity value satisfies a threshold value for merging agents, and wherein the instructions cause the processing circuitry to generate the second agent for the second persona based on the first source data and based further on the second source data. (GRINBERG − [0160] Some embodiments involve, when the calculated score is below a predetermined threshold, selecting an alternative AI agent from the pool of a plurality of AI agents, and transmitting the query to the alternative AI agent. A threshold, as used herein, may refer to a quantitative value or number that serves as a reference point that should be exceeded to proceed. A threshold may include a baseline, a limit (e.g., a maximum or minimum), a tolerance, a starting point, and/or an end point for a measurable quantity. A predetermined threshold may be a standard, goal, calculation, requirement, or suggestion. In some embodiments, the predetermined threshold may be preselected and stored in the system (e.g., during setup).)
Regarding dependent claim 9, depends on claim 8, GRINBERG teaches: wherein the instructions further cause the processing circuitry to: determine that each change of a second plurality of changes to a second plurality of historical documents is associated with one or more users and that each user of the one or more users is associated with the second persona, wherein the instructions cause the processing circuitry to generate the second agent for the second persona based on the first source data and the second source data and based further on the second plurality of changes. (GRINBERG − [0050] Examples of metadata may include user profiles, roles and permissions, activity logs, usage indications, preferences and settings, user associations/relationships, user history or a combination thereof. [0151] An AI agent, context, or a combination may be stored in a repository for use in an application. The combination of AI agent and context or query may have previously been determined to be a successful match. Consistent with some embodiments, selecting the particular AI agent from the pool of a plurality of AI agents includes accessing historical data on an ability of the particular AI agent to provide satisfactory responses in a context similar to the determined context. Historical data, as used herein, refers to any information regarding previous combinations of contexts, queries, and AI agents. Historical data may provide insights into the historical usage of an application, AI agent, user, or system.)
Regarding dependent claim 10, depends on claim 8, GRINBERG teaches: wherein the second persona is associated with a combined role; wherein the first persona is associated with a first portion of the combined role; and wherein the third persona is associated with a second portion of the combined role. (GRINBERG − [0151] A repository may include a list associating different AI agents with different contexts. For example, a first AI agent associated with questions related to places, a second AI agent associated with questions related to people, or some other association. A repository may include a set of various contexts or AI agents that may be used to address different or overlapping queries. An AI agent, context, or a combination may be stored in a repository for use in an application. The combination of AI agent and context or query may have previously been determined to be a successful match.)
Regarding dependent claim 11, depends on claim 1, GRINBERG teaches: wherein the instructions further cause the processing circuitry to: generate, based on the second agent and content of an electronic document, feedback information for the second persona; generate, based on the feedback information for the second persona, a multi-agent report for the electronic document; and output the multi-agent report. (GRINBERG – [0110] The AI agent may be configured to accept user feedback to improve performance. The AI agent may allow for additional operations, such as “summarize data,” in which the document may be analyzed and its contents summarized. Additional options from the AI agent may include “complete text,” in which the AI agent may complete a body of text that was started or initiated by a user input. [0141-0143] user interface to receive feedback)
Regarding dependent claim 12, depends on claim 1, GRINBERG teaches: wherein the instructions cause the processing circuitry to generate the first agent for the first persona based on both the first plurality of changes and first source data and based further on the second source data; and wherein to determine that the first source data is associated with the second persona, the instructions cause the processing circuitry to determine that the similarity value satisfies a threshold value for splitting agents. (GRINBERG − [0160] Some embodiments involve, when the calculated score is below a predetermined threshold, selecting an alternative AI agent from the pool of a plurality of AI agents, and transmitting the query to the alternative AI agent. A threshold, as used herein, may refer to a quantitative value or number that serves as a reference point that should be exceeded to proceed. A threshold may include a baseline, a limit (e.g., a maximum or minimum), a tolerance, a starting point, and/or an end point for a measurable quantity. A predetermined threshold may be a standard, goal, calculation, requirement, or suggestion. In some embodiments, the predetermined threshold may be preselected and stored in the system (e.g., during setup).)
Regarding dependent claim 13, depends on claim 12, GRINBERG teaches: wherein the instructions further cause the processing circuitry to: determine that each change of a second plurality of changes to a second plurality of historical documents is associated with one or more users and that each user of the one or more users is associated with the second persona, wherein the instructions cause the processing circuitry to generate the second agent for the second persona based on the first source data and based further on the second plurality of changes. (GRINBERG – [0059] Examples of metadata may include user profiles, roles and permissions, activity logs, usage indications, preferences and settings, user associations/relationships, user history or a combination thereof. [0103] Alternatively, the AI agents may be continually updated, modified, changed, or replaced, either by an existing AI agent, or by a developer. Selectable AI agents may include, for example, formula generators, planning generators, text generators, chat responders, and application finders. [0127] In this example, the predetermined AI agent may assess features such as the writing complexity, language, quality, and user history, when comparing the responses. [0143] [0137-0139] agent selection and updating based on user needs)
Regarding dependent claim 15, depends on claim 12, GRINBERG teaches: wherein the instructions further cause the processing circuitry to generate a third agent for a third persona based the second source data. (GRINBERG − [0160] Some embodiments involve, when the calculated score is below a predetermined threshold, selecting an alternative AI agent from the pool of a plurality of AI agents, and transmitting the query to the alternative AI agent. A threshold, as used herein, may refer to a quantitative value or number that serves as a reference point that should be exceeded to proceed. A threshold may include a baseline, a limit (e.g., a maximum or minimum), a tolerance, a starting point, and/or an end point for a measurable quantity. A predetermined threshold may be a standard, goal, calculation, requirement, or suggestion. In some embodiments, the predetermined threshold may be preselected and stored in the system (e.g., during setup).)
Regarding dependent claim 16, depends on claim 15, GRINBERG teaches: wherein the first persona is associated with a combined role; wherein the second persona is associated with a first portion of the combined role; and wherein the third persona is associated with a second portion of the combined role. (GRINBERG − [0160] Some embodiments involve, when the calculated score is below a predetermined threshold, selecting an alternative AI agent from the pool of a plurality of AI agents, and transmitting the query to the alternative AI agent. A threshold, as used herein, may refer to a quantitative value or number that serves as a reference point that should be exceeded to proceed. A threshold may include a baseline, a limit (e.g., a maximum or minimum), a tolerance, a starting point, and/or an end point for a measurable quantity. A predetermined threshold may be a standard, goal, calculation, requirement, or suggestion. In some embodiments, the predetermined threshold may be preselected and stored in the system (e.g., during setup).)
Regarding dependent claim 17, depends on claim 15, GRINBERG teaches: wherein the first persona is associated with a realtor role; wherein the second persona is associated with a buyer of the realtor role; and wherein the third persona is associated with a seller of the realtor role. (GRINBERG − [0065] In another example, different AI agents may be labelled by a specific type of data used in their training set, for example, a first AI agent may be specialized in customer relationships and trained to handle customer inquiries, a second AI agent may be specialized in financial analysis and have the ability to understand economic terminologies and analyze financial data,)
Regarding independent claim 18, is directed to a method. Claim 18 have similar/same technical features/limitations as claim 1 and the claims are rejected under the same rationale.
Regarding dependent claim 19, depends on claim 18, GRINBERG teaches: the method further comprising: generating, by the processing circuitry and based on the second agent and content of an electronic document, feedback information for the second persona; (GRINBERG – [0110] The AI agent may be configured to accept user feedback to improve performance. The AI agent may allow for additional operations, such as “summarize data,” in which the document may be analyzed and its contents summarized. Additional options from the AI agent may include “complete text,” in which the AI agent may complete a body of text that was started or initiated by a user input.)
generating, by the processing circuitry, data for a user interface configured to display an indication of the feedback information for the second persona; (GRINBERG – [0141-0143] Consistent with some embodiments, the feedback signal is received from a user via a user interface. [0143] Examples of allowing a user to provide feedback on a user interface are shown as thumbs up and thumbs down icons in FIG. 6 and FIG. 7.)
and outputting, by the processing circuitry and to a user device, the data for the user interface to cause the user device to display the user interface configured to display the indication of the feedback information for the second persona. (GRINBERG – [0140] Some disclosed embodiments involve receiving a feedback signal on the outputted response, wherein the feedback signal is either positive or negative; if the feedback signal is negative, selecting at least one other AI agent from the plurality of distinct AI agents; and outputting the response of the at least one other selected AI agent. A feedback signal in this context refers to a reaction to an outputted response. The reaction may be a response or information that is sent back to a system, process, or device associated with the output. The feedback signal is received when it is captured or obtained. In some embodiments, the signal may be used as a means for improvement, analysis, or decision making. [0143] Examples of allowing a user to provide feedback on a user interface are shown as thumbs up and thumbs down icons in FIG. 6 and FIG. 7.)
Regarding independent claim 20, is directed to a computer-readable media m. Claim 18 have similar/same technical features/limitations as claim 1 and the claims are rejected under the same rationale.
Claim(s) 14 is rejected under 35 U.S.C. 103 as being unpatentable over GRINBERG and Zheng as applied to claim 12 above, and further in view of Shen (US 20190171725 A1, Filed Date: Dec. 4, 2017).
Regarding dependent claim 14, depends on claim 12, GRINBERG does not explicitly teach: teaches: determine a centroid position for cluster of vector embeddings
Shen teaches: the instructions further cause the processing circuitry to: determine a centroid position for cluster of vector embeddings for the first agent, the cluster of vector embeddings for the first agent comprising the first plurality of vector embeddings and the second plurality of vector embeddings; and based on disturbing the centroid position of the cluster of vector embeddings, select a first subset of the cluster of vector embeddings as the first plurality of vector embeddings and a second subset of the cluster of vector embeddings as the second plurality of vector embeddings. (Shen − [0019] User profiles (P.sub.t and P.sub.t−1), relevant documents, relevant(t), and non-relevant documents, nonrelevant(t)), are all modeled as vectors in the same concept space. Relevant documents are those for which users showed interest (e.g. clicked documents), while non-relevant documents are those which users skipped. Mean(relevant(t)) and Mean(nonrelevant(t)) are the centroids of the relevant document vectors and non-relevant document vectors, respectively. [0084] FIG. 3 conceptually illustrates determination of a user profile vector in a vector space, in accordance with implementations of the disclosure. In the illustrated implementation, a vector space 300 is shown, in which content item vectors are defined. During a current time period t, a given user exhibits positive interactions with content items having vector representations A.sub.1, A.sub.2, A.sub.3, and A.sub.4, and the given user exhibits negative interactions with content items having vector representations A.sub.5, A.sub.6, A.sub.7, and A.sub.8. The positive interaction vector P.sub.t.sup.+ is determined as the centroid of the content item vectors A.sub.1, A.sub.2, A.sub.3, and A.sub.4. The negative interaction vector P.sub.t.sup.− is determined as the centroid of the content item vectors A.sub.5, A.sub.6, A.sub.7, and A.sub.8.)
Accordingly, it would have been obvious to one of ordinary skill in the art before effective filing date of the claim invention to have to incorporate the centroid-based representation taught by Shen into the vector embedding comparison techniques of Zhen when applied to GRINBERG’s agent framework. One would have been motivated to make such combination to improve the accuracy and consistency of GRINBERG’s comparison process when selecting an appropriate AI agent.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Baeum US 20230074406, LLM in generating automated AI responses.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to CARL E BARNES JR whose telephone number is (571)270-3395. The examiner can normally be reached Monday-Friday 9am-6pm.
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, Stephen Hong can be reached at (571) 272-4124. 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.
/CARL E BARNES JR/Examiner, Art Unit 2178
/STEPHEN S HONG/Supervisory Patent Examiner, Art Unit 2178
Prosecution Insights