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 .
Claim status
Claims 1-20 presented for the examination and remain pending in the application.
Claim objection
Claim 18 is objected to because of the following informalities: Claim 18 recites the term like “LLM” in line 1. This acronyms term is not defined in the claim at least for the first time. In regard to the term, “LLM” any abbreviation or shortened representations of a term should be spelled out completely upon its first use in each claim branch (e.g., user equipment (UE)). However, for the examination purpose, the Examiner has considered the acronyms term “LLM” in light of the Applicant’s Specification para. [0005]-[0006] as “Large Language Models”.
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 text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
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.
Claims 1-7, 9-15, 19 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Ahmadidaneshashtiani et al. U.S. Pub. No. 2021/0249002 A1, (hereinafter Ahmadid) in view of Dahan et al. U.S. Pub. No. 2016/0217159 A1, (hereinafter Dahan).
Regarding claim 1.
Ahmadid teaches a method for directing a domain-specific request to a machine learning (ML) agent of a plurality of ML agents configured to generate domain specific language (DSL) scripts (Ahmadid teaches in para. [0028] natural language processing agent machine learning (ML) and further, Ahmadid teaches in para. [0085] a domain-specific natural language (DSL) processing agent of the natural language processing agents 110 or 112 for intent detection, the selection based on a type of inquiry extracted from the input strings…, and further, Ahmadid teaches in para. [0134] particularly important where the public natural language processing agents utilize the utterances for training internal machine learning models, as it is not desirable to forge associations between sensitive information being used in utterances), the method comprising, at a computing system: processing a request received from a user to determine a particular domain associated with the request (Ahmadid teaches in para. [0085] and further, Ahmadid teaches in para. [0154] orchestrator system is adapted to request the user's profile information or a representation thereof from a client profile AI and insight engine…);
identifying an ML agent of the plurality of ML agents that is associated with the particular domain (Ahmadid teaches in para. [0039] select a domain-specific natural language processing engine of the natural language processing engines for intent detection, the selection based on a type of inquiry extracted from the input strings supported by machine learning algorithms that leverage personalization services);
generating, using the ML agent, a DSL script for the particular domain based on the request (Ahmadid teaches in para. [0097] Downstream APIs are managed in re-usable scripts referred to as handlers 114 that can be deployed as stand-alone services or copied into the broker service);
executing, using the ML agent, at least one of an API call and a database query based on the DSL script (Ahmadid teaches in para. [0097] the system 100 acts as a conversational agent (intermediary) that interoperates between an application interface, a NLP (either cloud based or any API enabled NLP), and downstream fulfilment APIs as illustrated in FIG. 1A. Downstream APIs are managed in re-usable scripts referred to as handlers 114 that can be deployed as stand-alone services or copied into the broker service and further, Ahmadid teaches in para. [0420] a server can include one or more computers operating as a web server, database server, or other type of computer server in a manner to fulfill described roles, responsibilities, or functions);
generating, using the ML agent, an output based on the response (Ahmadid teaches in para. [0138] generated using a machine learning model that establishes the next best word until a response is complete. The output response data structure encapsulates this response, and can be provided to the chatbot for downstream output); and
processing the output using at least one other ML agent of the plurality of ML agents (Ahmadid teaches in para. [0148] the system can be configured such that there may be multiple driving agents, each being associated to a corresponding user (e.g., in a multi-party example), or multiple instances of the orchestrator being run in parallel,..., and further, Ahmadid teaches in para. [0138] a string encapsulating the response from the agent, for example, generated using a machine learning model that establishes the next best word until a response is complete).
Ahmadid does not explicitly teach processing, using the ML agent, a response to the at least one of the API call and the database query.
However, Dahan teaches processing, using the ML agent, a response to the at least one of the API call and the database query (Dahan teaches in para. [0098] connection data, such as connection object 515a can be utilized throughout the application's 240 particular connection with the database 245, including queries of the database 245 made by the application 240, the application 240 can use the connection to query 525a the database…, and received connection object 510). The query method can be passed to the database API and the API, in some instances, can create a SQL or other statement or stored procedure as well as complete the statement by binding variables, setting optional attributes, etc.).
Therefore, Ahmadid and Dahan are analogues arts and they are in the same field of endeavor as they both are directed to machine learning and domain specific language to manage the database query.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the
claimed invention to modify the teachings of performing a particular connection including queries of the database and the query method can be passed to the database API and the API, in some instances, can create a SQL or other statement or stored procedure as well as complete the statement by binding variables, setting optional attributes, etc. ([0098]) as taught, by Dahnan into the teachings of Ahmadid invention. One would have been motivated to do so in order to the robust modeling of the variety of different session flows that can take place in system is allowed and the virtual services instantiated is allowed from the service model to be able to effectively simulate responses of the database that is received by live applications interfacing with the virtual service.
Regarding claim 2.
Ahmadid in view of Dahan teaches wherein the request is processed by another ML agent to determine the particular domain associated with the request before transmitting the request to the orchestrator (Ahmadid teaches in para. [0033] fallback utterance (e.g., please talk to human) is identified to be required by the orchestration server, and the computer processor is further configured to send machine-generated inputs to the natural language processing before responding to the client and further, Dahan teaches in para. [0108] a client 615 can interact with a frontend of a web service that causes the frontend server 605 to send a request to backend server 625. The backend server 625, upon receiving the request from the frontend server 605 may attempt to obtain a connection (e.g., a new connection or connection from connection pool 660) to the database (e.g., 635) to query the database and generate a response to the request of the frontend server 605).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of interacting with a frontend of a web service that causes the frontend server to send a request to backend server ([0108]) as taught, by Dahnan into the teachings of Ahmadid invention. One would have been motivated to do so in order to the virtual services provides functionality beyond traditional piecemeal responders or stubs, through logic permitting the recognition of input/requests and generation of outputs/responses that are stateful, aware of time, date, and latency characteristics.
Regarding claim 3.
Ahmadid teaches wherein the request is based on a natural language prompt received via a user interface (Ahmadid teaches in para. [0110] the system 100 is an intermediary that is configured to broker the conversation among an application interface, an NLU agent, and a series of fulfilment APIs. The system 100 acts as a conversational agent to nurture and maintain characteristics of a natural language conversation. Also, see para. [0127], [0363] and [0393]).
Regarding claim 4.
Ahmadid teaches wherein the request comprises a natural language prompt generated by another ML agent of the plurality of ML agents (Ahmadid teaches in para. [0148] the system can be configured such that there may be multiple driving agents, each being associated to a corresponding user (e.g., in a multi-party example), or multiple instances of the orchestrator being run in parallel,..., and further, Ahmadid teaches in para. [0138] a string encapsulating the response from the agent, for example, generated using a machine learning model that establishes the next best word until a response is complete).
Regarding claim 5.
Ahmadid teaches wherein the request comprises at least one of text data, image data, and audio data (Ahmadid teaches in para. [0016] utterance string inputs, in some embodiments, can include string inputs that are provided by other systems and not necessarily a human, and may be received the form of text, audio, voice, among others. Note that here, the claim lists features in the alternative. While the claim lists a number of optional limitations only one limitation from the list is required and needs to be met by the prior art. However, the prior art of record Ahmadid addressed all of the limitations).
Regarding claim 6.
Ahmadid teaches wherein the orchestrator is configured to determine the appropriate ML agent by comparing the particular domain associated with the request to a DSL definition of the ML agent (Ahmadid teaches in para. [0085] a domain-specific natural language (DSL) processing agent of the natural language processing agents 110 or 112 for intent detection, the selection based on a type of inquiry extracted from the input strings…, and further, Ahmadid teaches in para. [0313] one last advanced use case enabled by decoupling NLU's from the core experience is the ability to compare performance in a particular domain).
Regarding claim 7.
Ahmadid teaches wherein the DSL definition of the ML agent comprises an indication of an ability to generate a DSL script for the particular domain associated with the request (Ahmadid teaches in para. [0097] Downstream APIs are managed in re-usable scripts referred to as handlers 114 that can be deployed as stand-alone services or copied into the broker service).
Regarding claim 9.
Ahmadid teaches wherein the orchestrator is configured to dynamically utilize server-side computing resources, client-side computing resources, or a combination thereof based on an availability of computing resources (Ahmadid teaches in para. [0033] fallback utterance (e.g., please talk to human) is identified to be required by the orchestration server, and the computer processor is further configured to send machine-generated inputs to the natural language processing before responding to the client).
Regarding claim 10.
Ahmadid in view of Dahan teaches wherein any one or more of the plurality of ML agents is configured to dynamically utilize server-side computing resources, client-side computing resources, or a combination thereof based on an availability of computing resources (Ahmadid teaches in para. [0033] fallback utterance (e.g., please talk to human) is identified to be required by the orchestration server, and the computer processor is further configured to send machine-generated inputs to the natural language processing before responding to the client and further, Dahan teaches in para. [0104] a live session can be observed and tracked using agents 670, 680 instrumented on software components in each of front-end server 605 and back-end server 625…, and further, Dahan teaches in para. [0106] an additional request may be generated by the client 615 in a new set of transactions 695 that is to result in a query of a database by a backend server supporting the web service…).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of a live session can be observed and tracked using agents instrumented on software components in each of the frontend server to send a request to backend server ([0104]) as taught, by Dahnan into the teachings of Ahmadid invention. One would have been motivated to do so in order to the virtual services provides functionality beyond traditional piecemeal responders or stubs, through logic permitting the recognition of input/requests and generation of outputs/responses that are stateful, aware of time, date, and latency characteristics.
Regarding claim 11.
Ahmadid teaches wherein the output is transmitted to a user interface ML agent trained to reconfigure a user interface based on the output (Ahmadid teaches in para. [0393] the system (100) connects to human response interfaces for feedback. The computer processor is further configured to process the feedback and send machine-generated inputs to the natural language processing for further instructions or intent data objects. This implementation creates a multi-agent conversation which is superior to machine-to-machine communications as it is easily understandable by humans and the dialog systems).
Regarding claim 12.
Ahmadid teaches receiving, by the user interface ML agent, a user input via the user interface (Ahmadid teaches in para. [0037] the system includes a computer processor operating in conjunction with computer memory. The computer processor is configured to receive, from a user interface (e.g., an audio or a visual interface rendered on a display of a mobile device or an automated teller machine (ATM)), input strings representing a conversational request from the user having query tokens); and
reconfiguring the user interface based on the user input (Ahmadid teaches in para. [0024] the orchestration agent can be configured to automatically conduct a verification step where the interface is controlled to ask the user which domain or agent specialization it is expecting to utilize…, the orchestration system may be controlled to initiate a verification step where the user is prompted with a decision input request asking which of the topics of wealth management and personal banking is a better fit for supporting the conversation).
Regarding claim 13.
Ahmadid in view of wherein reconfiguring the user interface comprises: selecting one or more widgets based on the user input and data included in the response from at least one of the API and the database (Ahmadid teaches in para. [0420] a server can include one or more computers operating as a web server, database server, or other type of computer server in a manner to fulfill described roles, responsibilities, or functions and further, Dahan teaches in para. [0098] connection Dahan teaches data, such as connection object 515a can be utilized throughout the application's 240 particular connection with the database 245, including queries of the database 245 made by the application 240, the application 240 can use the connection to query 525a the database).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the
claimed invention to modify the teachings of performing a particular connection including queries of the database and the query method can be passed to the database API and the API, in some instances, can create a SQL or other statement or stored procedure as well as complete the statement by binding variables, setting optional attributes, etc. ([0098]) as taught, by Dahnan invention. One would have been motivated to do so in order to the robust modeling of the variety of different session flows that can take place in system is allowed and the virtual services instantiated is allowed from the service model to be able to effectively simulate responses of the database that is received by live applications interfacing with the virtual service.
Regarding claim 14.
Ahmadid teaches wherein a DSL interpreter interprets the DSL script during runtime to execute at least one of an API call and a database query based on the DSL script (Ahmadid teaches in para. [0387] the fulfilment handlers call and execute the backend services, the fulfilment handlers can re-verify authorization and credentials for a given user at the desired control point in the conversational flow. The fulfillment handler engines route communications to a plurality of API endpoints for a plurality of outbound channels and in some embodiments, the conversational middleware platform 110 is configured to select between an API-based conversation and a socket-based conversation base at least on a period of time).
Regarding claim 15.
Ahmadid teaches wherein the DSL interpreter is included in the ML agent (Ahmadid teaches in para. [0032]-[0033] a fulfillment string can be generated for invoking a function to conduct the fulfillment (e.g., a command line interpreter instruction message)…, fallback utterance (e.g., please talk to human) is identified to be required by the orchestration server, and the computer processor is further configured to send machine-generated inputs to the natural language processing before responding to the client).
Regarding claims 19 and 20.
Claims 19 and 20 incorporate substantively all the limitation of claim 1 in a computer program product and a system form and are rejected under the same rationale. Furthermore, regarding the limitations of a non-transitory computer readable storage medium, DSL and machine learning, the prior art of record Ahmadid teaches in para. [0009], [0028] and [0085].
Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Ahmadid in view of Dahan further in view of Bhagwan et al. U.S. Pub. No. 2023/0267178 A1, (hereinafter Bhagwan).
Regarding claim 8. Ahmadid in view of Dahan teaches the method of claim 6.
Ahmadid in view of Dahan does not explicitly teach wherein the DSL definition of the ML agent is specified via a manifest file definition of the ML agent.
However, Bhagwan teaches wherein the DSL definition of the ML agent is specified via a manifest file definition of the ML agent (Bhagwan teaches in para. [0035] rule learning engine 320 may utilize any pattern matching software, such as machine learning based pattern matching, or string profiling 325…, and further, Bhagwan teaches in para. [0036] rule learning engine 320 provides the learned rules to a rule database 335, which stores the rules. The rules may be stored in a format that includes a file name 340,…).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the
claimed invention to modify the teachings of rule based rules storing in a format that includes a file name and the system provides the learned rules to a rule database ([0035]-[0036]) as taught, by Bhagwan into the teachings of Ahmadid in view of Dahnan invention. One would have been motivated to do so in order to the robust modeling of the variety of different session flows that can take place in system is allowed and the virtual services instantiated is allowed from the service model to be able to effectively simulate responses of the database that is received by live applications interfacing with the virtual service.
Claims 16 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Ahmadid in view of Dahan further in view of Siebel et al. U.S. Pub. No. 2024/0202225 A1, (hereinafter Siebel).
Regarding claim 16. Ahmadid in view of Dahan teaches the method of claim 1.
Ahmadid in view of Dahan does not explicitly teach wherein the output is transmitted to each of the plurality of ML agents for validating the output.
However, Siebel teaches wherein the output is transmitted to each of the plurality of ML agents for validating the output (Siebel teaches in para. [0123] the artificial intelligence traceability module 516 can provide sources citations automatically and/or on-demand to corroborate or validate large language model outputs. The artificial intelligence traceability module 516 may also determine a compatibility of the different sources (e.g., data records, passages) that were used to generate a large language model output…).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the
claimed invention to modify the teachings of providing sources citations automatically and/or on-demand to corroborate or validate large language model outputs ([0123]) as taught, by Siebel into the teachings of Ahmadid in view of Dahnan invention. One would have been motivated to do so in order to the enterprise generative artificial intelligence system efficiently provides more accurate and reliable results than conventional generative artificial intelligence solutions when consuming fewer computing resources and requiring shorter processing times. The method prevents the large language model from hallucinating and allows the large language model to provide an accurate and reliable final result. The operation improves the computational efficiency of providing an output in response to an input.
Regarding claim 18. Ahmadid in view of Dahan teaches the method of claim 1.
Ahmadid in view of Dahan does not explicitly teach wherein each ML agent of the plurality of ML agents comprises an LLM trained for a different task than an LLM of each of the other ML agents of the plurality of ML agents.
However, Siebel teaches wherein each ML agent of the plurality of ML agents comprises an LLM trained for a different task than an LLM of each of the other ML agents of the plurality of ML agents (Siebel teaches in para. [0023] machine learning models can include some or all of the different types or modalities of models described herein (e.g., multimodal machine learning models, large language models (LLM), data models, statistical models, audio models, visual models, audiovisual models, etc.). Agents can include one or more multimodal models (e.g., large language models) to accomplish the prescribed tasks using a variety of different tools… Also, see para. [0024]-[0028]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the
claimed invention to modify the teachings of using machine learning models which can include LLM ([0023]-[0024]) as taught, by Siebel into the teachings of Ahmadid in view of Dahan invention. One would have been motivated to do so in order to accomplish the prescribed tasks using a variety of different tools. Different agents can use various tools to execute and process unstructured data retrieval requests, structured data retrieval requests, API calls (e.g., for accessing artificial intelligence application insights), and the like. Tools can include one or more specific functions and/or machine learning models to accomplish a given task (or set of tasks). (Siebel. [0023]-[0024]).
Claim 17 is rejected under 35 U.S.C. 103 as being unpatentable over Ahmadid in view of Dahan further in view of Siebel and further in view of Shete et al. U.S. Pub. No. 2023/0015083 A1, (hereinafter Shete).
Regarding claim 17. Ahmadid in view of Dahan further in view of Siebel teaches the method of claim 16.
Ahmadid in view of Dahan further in view of Siebel does not explicitly teach wherein validating the output comprises performing a consensus protocol by the plurality of ML agents based on pretrained knowledge of the subset of the plurality of ML agents and external data sources.
However, Shete teaches wherein validating the output comprises performing a consensus protocol by the plurality of ML agents based on pretrained knowledge of the subset of the plurality of ML agents and external data sources (Shete teaches in para. [0007] using pretrained Machine Learning (ML) models to detect one or more agents that will most likely accept one or more time-intervals from the identified one or more time-intervals to store the one or more agents in the database of a plurality of agents with respective plurality of scheduled-working-shifts…Also, see para. [0061], [0085] and [0087]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the
claimed invention to modify the teachings of using pretrained Machine Learning (ML) models ([0007]) as taught, by Shete into the teachings of Ahmadid in view of Dahan further in view of Siebel invention. One would have been motivated to do so in order to the method enables identifying time intervals being understaffed or overstaffed, and identifying and reaching out agents likely take advantage of extra hours or time-off opportunities with lesser turnaround time, so as to reduce time and resources consumption.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BERHANU SHITAYEWOLDETSADIK whose telephone number is (571)270-7142. The examiner can normally be reached M-F.
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, Emmanuel Moise can be reached at 5712723865. 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.
/BERHANU SHITAYEWOLDETSADIK/Primary Examiner, Art Unit 2455
