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 .
Response to Arguments
Applicant's arguments filed 2/5/26 have been fully considered but they are not persuasive.
Regarding the rejection of claims 7 and 8, Applicant argues that the subsequent Office Action (to correct the typo in the rejection) should not be made final despite Examiner explicitly identifying during the provided interview held 2/4/26 that the rejection was in view of Gutierrez (Arguments, pg. 8- pg. 9, second para.). Examiner respectfully disagrees as correcting a typo does not constitute new grounds of rejection. As mentioned during the interview and as presented in the interview summary record dated 2/6/26, the Office Action (12/2/26, pg. 2-4) under the 35 U.S.C. 102 rejection heading incorrectly lists the rejection of claims 7 and 8 as being rejected with Gutierrez in view of Weston instead of being rejected with just Gutierrez. Examiner mentioned this typo during the interview as admitted by Applicant in Applicant’s provided arguments, and given that Weston’s disclosure do not extend beyond paragraph [0123], while the citations present in the rejection of claims 7 and 8 (Non-Final rejection, 12/2/25, page 4) included paragraphs [0170], [0238], [0642],[0473] and [0500], it is unclear why this still remains an issue with Applicant seemingly suggesting that a correction to this already identified typo would constitute new grounds of rejection. The typo is hereby corrected.
Applicant further argues that the Office action appears to interpret Gutierrez’s verb schemas as the claimed one or more design artifacts and that Gutierrez describes Verb Schemas, Code Package Schemas, Mapping Schemas, and standard knowledge data stores (SKDSs) as distinct structures unlike the amended claim that requires claim 1 requires creation of a single digital assistant definition, and as such, argues that Gutierrez fails to disclose limitation “creating a digital assistant definition comprising one or more digital assistant design-time artifacts, wherein the one or more digital assistant design-time artifacts comprise a skills digital assistant design-time artifact, an intents digital assistant design-time artifact, and an entities digital assistant design-time artifact, wherein the given intent and the invocation logic definition for the given intent are integrated as linked in the digital assistant definition, wherein the digital assistant definition is consumable by a digital assistant compiler or interpreter to generate the executable digital assistant” as recited in claim 1 and as similarly recited in claims 17 and 20 (Arguments, pg. 9, third para. – pg. 10; pg. 11).
Examiner respectfully disagrees as Gutierrez discloses automatically generating Schemas from APIs and documentation where a modifiable PRE/machine learning model/virtual assistant/chatbot is trained, tuned and modified with the schemas (para. [0500]; para. [0635]; para. [0637]; para. [0649]), and where each Schema includes nouns (i.e., the claimed entities), Verbs (i.e., claimed intents), mappings (i.e., claimed skills) and integrations/capabilities (i.e. the claimed invocation definition) see paragraphs [0185], [0238]; para. [0258]; and para. [0357]), and where the verbs/intents are linked/mapped with the integrations/capabilities (para. [0185]; para. [0246]; para. [0357]), corresponding to limitation “creating a digital assistant definition comprising one or more digital assistant design-time artifacts, wherein the one or more digital assistant design-time artifacts comprise a skills digital assistant design-time artifact, an intents digital assistant design-time artifact, and an entities digital assistant design-time artifact, wherein the given intent and the invocation logic definition for the given intent are integrated as linked in the digital assistant definition, wherein the digital assistant definition is consumable by a digital assistant compiler or interpreter to generate the executable digital assistant”.
Regarding the 35 U.S.C. 103 rejection of the dependent claims 2-6, 10-14, 18 and 18 with additional references Weston, Rennie and Mishra, Applicant argues that the additional references do not remedy the above argued deficiencies of Gutierrez, and as such the additional references do not disclose language recited in the dependent claims (Arguments, pg. 11)
Examiner respectfully disagrees as the additional references were/are not utilized to teach the above argued limitation, and absent any argument as to why the cited portions of the references fail to address limitations of the dependent claims, Examiner maintains that the rejections of the dependent claims are appropriate.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
1. Claims 1, 6-9, 15 and 16 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Gutierrez US 2024/0419706 A1 (“Gutierrez”)
Per claim 1, Gutierrez discloses a computer-implemented method for automated generation of an executable digital assistant, the method comprising:
loading one or more large language models with one or more input documents describing functionality of a software application (para. [0442]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]);
prompting at least one of the one or more large language models to provide a list of possible intents that can be performed by users in the software application (certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0142]-[0143]; the types of Schemas that define extensible, and often standardized, abstractions of software capabilities in SKL may be called “Verbs.” In other words, a “Verb” may provide a “standardized” representation of a certain software process or capability offered by one or more software tools …, para. [0163]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500], generated schemas as including verbs/capability, software capabilities/verbs as intents);
determining an invocation logic definition for a given intent out of the list of possible intents (In addition to a Verb Schema, a Mapping (defined below) Schema may determine what will happen when a Verb is called …, para. [0170]-[0177]; an “Integration” may be a type of Schema, and more specifically, a type of Noun that may represent the data, capabilities …, para.[0185]; According to some embodiments, a VERBINTEGRATIONMAPPING may translate the inputs of a Verb to the unique inputs and correct capability (API endpoint, SDK function call, etc.) of an Integration to execute and perform the intent of the Verb using the Integration …, para. [0246]; there may be an abstraction/translation of software and components (e.g., proprietary APIs and their data) into a universal language of data, capabilities …, para. [0332]; para. [0357]; para. [0380]; para. [0431]); and
creating a digital assistant definition comprising one or more digital assistant design-time artifacts, wherein the given intent and the invocation logic definition for the given intent are integrated as linked in the digital assistant definition (the Verbs available to an application using a Standard SDK (defined below), or to an end user of such an application, may be defined via Schemas. Each Verb Schema may define Metadata about the Verb such as its name and description, as well as its standard inputs and outputs. In addition to a Verb Schema, a Mapping (defined below) Schema may determine what will happen when a Verb is called …, para. [0170]-[0177]; para. [0185]; para. [0191]; para. [0199]; each Schema representing a Noun, Verb, Integration, Code Package, Interface, and/or SKDS may have embedded within it the logic and translations for how it relates to, uses, or produces every other artifact…, para. [0238]; para. [0242]-[0245]; According to some embodiments, a VERBINTEGRATIONMAPPING may translate the inputs of a Verb to the unique inputs and correct capability …, para. [0246]; para. [0257]-[0263]; para. [0332]; para. [0357]; para. [0366]; para. [0444]; para. [0500]; When the user or developer wants to support an additional integration, all he/she may need to do is add an additional translation/Mapping between the standard Verb and the new Integration …, para. [0551]; para. [0737], verbs/intents mapped to integrations/capabilities);
creating a digital assistant definition comprising one or more digital assistant design-time artifacts, wherein the one or more digital assistant design-time artifacts comprise a skills digital assistant design-time artifact, an intents digital assistant design-time artifact, and an entities digital assistant design-time artifact, wherein the given intent and the invocation logic definition for the given intent are integrated as linked in the digital assistant definition, wherein the digital assistant definition is consumable by a digital assistant compiler or interpreter to generate the executable digital assistant (para. [0022]; para. [0169]-[0170]; para.[0185]; VerbIntegrationMapping: …, para. [0246]; A Standard Knowledge Data Store or “SKDS” may be a type of Integration that stores Entities, and in some embodiments, also stores the Schemas that make up the Nouns, Verbs, and Mappings and/or other SKL components on behalf of a developer …, para. [0258]; para. [0332]; para. [0357]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]; the SKL systems and methods described herein may be implemented into a computing infrastructure that includes one or more artificial intelligence (“AI”) and/or machine-learning models/architectures for processing a natural language user query received by a user device … An analytics server executing a machine-learning model (referred to herein as a Personalized Response Engine (“PRE”)) may utilize the schema-based SKL components described herein to generate one or more personalized elements (e.g., responses) to present to the user based on the original user query …, para. [0635]; the user may converse with the PRE when implemented as a chatbot …, para. [0637]; para. [0649], nouns, verbs and mappings as entity, intent and skills artifacts, generated schema including nouns, verbs, mappings and integration mappings as integrated as part of schema mappings, generated and modified chatbot/assistant/PRE as utilizing schema of nouns, verbs, mappings and integrations/capabilities, implementing the schema into a modifiable PRE/machine learning model/virtual assistant/chatbot as implying use of compiler or interpreter).
Per claim 6, Gutierrez discloses the method of claim 1, wherein: creating the digital assistant definition comprises: creating the intents digital assistant design-time artifact comprising the given intent (para. [0238]; para. [0242]-[0246]; para. [0258]; para. [0262]; para. [0500]; para. [0737]); and
creating the skills digital assistant design-time artifact comprising the invocation logic definition (para. [0161]; para. [0258]; para. [0500]).
Per claim 7, Gutierrez discloses the method of claim 1, wherein: creating the digital assistant definition comprises: creating the intents digital assistant design-time artifact comprising the given intent (para. [0170]; para. [0238]; para. [0258]; para. [0642]); and
integrating the invocation logic definition in the intents digital assistant design- time artifact (para. [0170]; each Schema representing a Noun, Verb, Integration, Code Package, Interface, and/or SKDS may have embedded within it the logic … Mappings defined separately from the Schemas, configurations, or code of Nouns, Verbs, Integrations …, para. [0238]; para. [0246]; para. [0258]; para. [0357]; para. [0500]).
Per claim 8, Gutierrez discloses the method of claim 1, wherein: the input documents describe functionality of a suite of software applications (para. [0473]; According to a non-limiting embodiment, an analytics server could automatically generate Schemas from APIs and/or Documentation …, para. [0500]).
Per claim 9, Gutierrez discloses the method of claim 1, wherein: determining the invocation logic definition for the given intent comprises: converting an API definition into a natural language format (there may be an abstraction/translation of software and components (e.g., proprietary APIs and their data) into a universal language of data, capabilities … the various components are represented abstractly through Schemas which include Nouns 1920, Verbs 1930, Integrations …, para. [0332]; para. [0351]; para. [0473]; The API documentation generated could be in the format of an open-source API documentation standard such as OpenAPI in YAML, XML, or JSON …, para. [0498]-[0499]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources …, para. [0500]; para. [0505]; para. [0631]; para. [0684]);
loading at least one of the large language models with the natural language format (para. [0500]); and
prompting the at least one of the large language models to generate the invocation logic definition for the given intent (para. [0500]).
Per claim 15, Gutierrez discloses the method of claim 1, further comprising: deploying a digital assistant according to the digital assistant definition (para. [0258]; para. [0500]; para. [0635]);
receiving user prompts with the digital assistant (para. [0052]-[0053]; para. [0635]); and
outputting answers to the user prompts with the digital assistant (para. [0052]-[0053]; para. [0635]).
Per claim 16, Gutierrez discloses the method of claim 1, wherein: determining the invocation logic definition comprises: training the one or more large language models with examples of invocation logic definitions associated with respective intents (para. [0185]; para. [0357]; certain machine-learning models could be trained and/or tuned with SKL Schemas in order to help with the creation and maintenance of said Schemas. …, para. [0500]); and
requesting the one or more large language models to generate the invocation logic definition for the given intent (para. [0176]; para. [0185]; para. [0357]; para. [0500]).
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
2. Claims 2-6 and 17-19 are rejected under 35 U.S.C. 103 as being unpatentable over Gutierrez in view of Weston et al US 2025/0054493 A1 (“Weston”)
Per claim 2, Gutierrez discloses the method of claim 1,
Gutierrez does not explicitly disclose for the given intent out of the possible intents, prompting at least one of the one or more large language models to generate a plurality of exemplar intent utterances that users would provide as prompts to perform the given intent or integrating the exemplar intent utterances into the digital assistant definition
However, these features are taught by Weston:
for the given intent out of the possible intents, prompting at least one of the one or more large language models to generate a plurality of exemplar intent utterances that users would provide as prompts to perform the given intent para. [0036]; the electronic device 14 may generate a list of training utterances for each intent in the list of intents. Generating the list of training utterances may occur automatically upon generating the list of intents. Each training utterance may be associated with an intent …, para. [0083]; the training utterances may be generated by a generative pretrained transformer (GPT) prompted to generate one or more utterances that include the intent …, para. [0091]); and
integrating the exemplar intent utterances into the digital assistant definition (para. [0021]; para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0091]-[0093]; para. [0116])
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Weston with the method of Gutierrez in arriving at the missing features of Gutierrez, because such combination would have resulted in improving the accuracy or completeness of classifications with respect to input utterances (Weston, para. [0100])
Per claim 3, Gutierrez in view of Weston discloses the method of claim 2,
Weston discloses prompting at least one of the one or more large language models to identify one or more entities in the exemplar intent utterances (fig. 5; Intent determination typically employs techniques such as named entity recognition (NER) …, para. [0003]; para. [0087]; Once the NLU model is trained based on the algorithm and training data set, the NLU model may be used to classify unlabeled data points into known classifications (e.g., domains and/or intents) …, para. [0101]; Block 510 may be iterative, meaning it may be run several times, tweaking various parameters, thereby annealing the NLU model until it performs satisfactorily.…, para. [0103]-[0104]; The NLU model may be trained based on a list of training utterances …, para. [0105]-[0110]); and
integrating the entities into the digital assistant definition (para. [0003]; para. [0021]; para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0091]-[0093]; para. [0116])
Per claim 4, Gutierrez in view of Weston discloses the method of claim 3,
Gutierrez discloses integrating the entities into the digital assistant definition comprises integrating the entities into the entities digital assistant design-time artifact (para. [0258]; para. [0500])
Weston discloses integrating the entities into the digital assistant definition comprises integrating the entities into the entities digital assistant design-time artifact (Intent determination typically employs techniques such as named entity recognition (NER) …, para. [0003]; para. [0021]; para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0086]; para. [0091]-[0093]; Block 510 may be iterative, meaning it may be run several times, tweaking various parameters, thereby annealing the NLU model until it performs satisfactorily.…, para. [0103]-[0104]; The NLU model may be trained based on a list of training utterances …, para. [0105]-[0110]; para. [0116], entities as present in generated utterances).
Per claim 5, Gutierrez in view of Weston discloses the method of claim 2,
Weston discloses wherein: integrating the exemplar intent utterances into the digital assistant definition comprises integrating the exemplar intent utterances into the intents digital assistant design-time artifact (para. [0021]; para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0091]-[0093]; para. [0116]).
Per claim 17, Gutierrez discloses a computing system for automated generation of an executable digital assistant, the method comprising:
at least one hardware processor (para. [0020]);
at least one memory coupled to the at least one hardware processor (para. [0020]),
a stored internal representation of one or more large language models (para. [0682]; para. [0893]-[0894]), and
a digital assistant generation orchestrator configured to accept a plurality of documents describing a suite of one or more software applications (para. [0442]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]),
submit at least one of the documents to at least one of the large language models as a learning document, prompt the at least one of the large language models for a list of intents (certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0142]-[0143]; the types of Schemas that define extensible, and often standardized, abstractions of software capabilities in SKL may be called “Verbs.” In other words, a “Verb” may provide a “standardized” representation of a certain software process or capability offered by one or more software tools …, para. [0163]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500], generated schemas as including verbs/capability, software capabilities/verbs as intents),
prioritize the intents (para. [0380]; In another example, a user of an SKL Library, such as a developer or an application, may provide certain parameters, such as certain integrations, nouns, verbs, interface, etc. of interest, to the Library or Libraries in order to get a packaged set of Schemas back from the Library or Libraries that best match the capabilities that the user specified.…, para. [0631]),
output a digital assistant definition comprising one or more digital assistant design-time artifacts comprising at least the given intent, wherein the one or more digital assistant design-time artifacts comprise a skills digital assistant design-time artifact, an intents digital assistant design-time artifact, and an entities digital assistant design-time artifact, wherein the digital assistant definition is consumable by a digital assistant compiler or interpreter to generate the executable digital assistant (para. [0022]; para. [0169]-[0170]; para.[0185]; para. [0238]; para. [0246]; A Standard Knowledge Data Store or “SKDS” may be a type of Integration that stores Entities, and in some embodiments, also stores the Schemas that make up the Nouns, Verbs, and Mappings and/or other SKL components on behalf of a developer …, para. [0258]; para. [0332]; para. [0357]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]; the SKL systems and methods described herein may be implemented into a computing infrastructure that includes one or more artificial intelligence (“AI”) and/or machine-learning models/architectures for processing a natural language user query received by a user device … An analytics server executing a machine-learning model (referred to herein as a Personalized Response Engine (“PRE”)) may utilize the schema-based SKL components described herein to generate one or more personalized elements (e.g., responses) to present to the user based on the original user query …, para. [0635]; the user may converse with the PRE when implemented as a chatbot …, para. [0637]; para. [0649], nouns, verbs and mappings as entity, intent and skills artifacts, generated schema including nouns, verbs, mappings and integration mappings as integrated as part of schema mappings, generated and modified chatbot/assistant/PRE as utilizing schema of nouns, verbs, mappings and integrations/capabilities, implementing the schema into a modifiable PRE/machine learning model/virtual assistant/chatbot as implying use of compiler or interpreter).
Gutierrez does not explicitly disclose submit at least one of the intents as a given intent to one or more of the large language models, prompt the one or more of the large language models for a list of utterances or output one or more digital assistant design-time artifacts comprising at least the one or more of the utterances
However, these features are taught by Weston:
submit at least one of the intents as a given intent to one or more of the large language models, prompt the one or more of the large language models for a list of utterances (para. [0036]; the electronic device 14 may generate a list of training utterances for each intent in the list of intents. Generating the list of training utterances may occur automatically upon generating the list of intents. Each training utterance may be associated with an intent …, para. [0083]; the training utterances may be generated by a generative pretrained transformer (GPT) prompted to generate one or more utterances that include the intent …, para. [0091]-[0093]), and
output one or more digital assistant design-time artifacts comprising at least one or more of the utterances (para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0091]-[0093])
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Weston with the method of Gutierrez in arriving at the missing features of Gutierrez, because such combination would have resulted in improving the accuracy or completeness of classifications with respect to input utterances (Weston, para. [0100]).
Per claim 18, Gutierrez in view of Weston discloses the system of claim 17,
Gutierrez discloses wherein: the digital assistant generation orchestrator is further configured to accept a document describing an API of the suite of one or more software applications, submit the document describing the API to one or more of the large language models, prompt the at least one of the large language models for a list of intents, and aggregate the intents of the document describing the API with other intents (para. [0442]; para. [0473]; para. [0500]).
Per claim 19, Gutierrez in view of Weston discloses the system of claim 17
Gutierrez discloses the digital assistant compiler configured to compile the one or more digital assistant design-time artifacts into the executable digital assistant (para. [0170]; each Schema representing a Noun, Verb, Integration, Code Package, Interface, and/or SKDS may have embedded within it the logic … Mappings defined separately from the Schemas, configurations, or code of Nouns, Verbs, Integrations …, para. [0238]; the SKL systems and methods described herein may be implemented into a computing infrastructure that includes one or more artificial intelligence (“AI”) and/or machine-learning models/architectures for processing a natural language user query … An analytics server executing a machine-learning model (referred to herein as a Personalized Response Engine (“PRE”)) may utilize the schema-based SKL components described herein to generate one or more personalized elements (e.g., responses) to present to the user based on the original user query …, para. [0635]; para. [0642]; para. [0647]; para. [0649], implementing the schema into a modifiable PRE/machine learning model/virtual assistant/chatbot as implying use of compiler or interpreter).
3. Claims 11-14 are rejected under 35 U.S.C. 103 as being unpatentable over Gutierrez in view of Rennie et al US 2024/0265041 A1 (“Rennie”)
Per claim 11, Gutierrez discloses the method of claim 1,
Gutierrez does not explicitly disclose determining top intents, wherein determining the top intents comprises aggregating lists of possible intents from different source documents or selecting the given intent from the top intents
However, these features are taught by Rennie
determining top intents, wherein determining the top intents comprises aggregating lists of possible intents from different source documents (para. [0026]-[0028]; para. [0033]; para. [0197]); and
selecting the given intent from the top intents (para. [0197])
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Rennie with the method of Gutierrez in arriving at the missing features of Gutierrez, because such combination would have resulted in improving the quality of searching operations during runtime (Rennie, para. [0005]).
Per claim 12, Gutierrez discloses the method of claim 1,
Gutierrez does not explicitly disclose wherein: at least one of the one or more large language models is of a different large language model type
However, this feature is taught by Rennie (para. [0049]; para. [0112]-[0124])
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Rennie with the method of Gutierrez in arriving at the missing features of Gutierrez, because such combination would have resulted in implementing different training and prediction schemes (Rennie, para. [0124]).
Per claim 13, Gutierrez discloses the method of claim 12,
Gutierrez does not explicitly disclose wherein: the at least one of the one or more large language models is tailored to a document type that it receives
However, this feature is taught by Rennie (para. [0049)
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Rennie with the method of Gutierrez in arriving at the missing features of Gutierrez, because such combination would have resulted in determining a document is to be analyzed and how a resultant edited content should be generated (Rennie, para. [0257]).
Per claim 14, Gutierrez in view of Rennie discloses the method of claim 12,
Gutierrez discloses wherein: the at least one of the one or more large language models is tailored to interpret process diagrams (para. [0479]; para. [0500]).
4. Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Gutierrez in view of Weston and Mishra et al US 12,456,020 B1 (“Mishra”)
Per claim 20, Gutierrez discloses one or more non-transitory computer-readable media comprising computer-executable instructions that, when executed by a computing system, cause the computing system to perform a method for automated generation of an executable digital assistant, the method comprising:
loading a first large language model with documentation of a software application (para. [0442]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]);
prompting the first large language model to provide a list of possible intents that can be performed by users in the software application (certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0142]-[0143]; the types of Schemas that define extensible, and often standardized, abstractions of software capabilities in SKL may be called “Verbs.” In other words, a “Verb” may provide a “standardized” representation of a certain software process or capability offered by one or more software tools …, para. [0163]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500], generated schemas as including verbs/capability, software capabilities/verbs as intents);
loading a large language model with documentation of APIs of the software application (para. [0442]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]);
prompting the second large language model to provide a list of possible intents that can be performed with the APIs of the software application (certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0142]-[0143]; the types of Schemas that define extensible, and often standardized, abstractions of software capabilities in SKL may be called “Verbs.” In other words, a “Verb” may provide a “standardized” representation of a certain software process or capability offered by one or more software tools …, para. [0163]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500], generated schemas as including verbs/capability, software capabilities/verbs as intents);
identifying at least one given intent out of the lists of possible intents (According to some embodiments, a VERBINTEGRATIONMAPPING may translate the inputs of a Verb to the unique inputs and correct capability (API endpoint, SDK function call, etc.) of an Integration to execute and perform the intent of the Verb using the Integration …, para. [0246]; para. [0380]);
determining an API invocation logic definition for the given intent (According to some embodiments, a VERBINTEGRATIONMAPPING may translate the inputs of a Verb to the unique inputs and correct capability (API endpoint, SDK function call, etc.) of an Integration to execute and perform the intent of the Verb using the Integration …, para. [0246]; para. [0380]); and
creating a digital assistant definition with the given intent, the API invocation logic definition, wherein the digital assistant definition comprises one or more digital assistant design-time artifacts, and wherein the one or more digital assistant design-time artifacts comprise a skills digital assistant design-time artifact, an intents digital assistant design-time artifact, and an entities digital assistant design-time artifact, wherein the digital assistant definition is consumable by a digital assistant compiler or interpreter to generate the executable digital assistant (para. [0022]; para. [0169]-[0170]; para.[0185]; VerbIntegrationMapping: …, para. [0246]; A Standard Knowledge Data Store or “SKDS” may be a type of Integration that stores Entities, and in some embodiments, also stores the Schemas that make up the Nouns, Verbs, and Mappings and/or other SKL components on behalf of a developer …, para. [0258]; para. [0332]; para. [0357]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources (e.g., documentation about APIs …, para. [0500]; the SKL systems and methods described herein may be implemented into a computing infrastructure that includes one or more artificial intelligence (“AI”) and/or machine-learning models/architectures for processing a natural language user query received by a user device … An analytics server executing a machine-learning model (referred to herein as a Personalized Response Engine (“PRE”)) may utilize the schema-based SKL components described herein to generate one or more personalized elements (e.g., responses) to present to the user based on the original user query …, para. [0635]; the user may converse with the PRE when implemented as a chatbot …, para. [0637]; para. [0649], nouns, verbs and mappings as entity, intent and skills artifacts, generated schema including nouns, verbs, mappings and integration mappings as integrated as part of schema mappings, generated and modified chatbot/assistant/PRE as utilizing schema of nouns, verbs, mappings and integrations/capabilities, implementing the schema into a modifiable PRE/machine learning model/virtual assistant/chatbot as implying use of compiler or interpreter).
Gutierrez does not explicitly disclose for the given intent, prompting a large language model to generate a plurality of exemplar utterances that users could provide to perform the given intent, wherein the exemplar utterances comprise indications of entities, storing the given intent in a design-time format for a digital assistant or creating a digital assistant definition with the exemplar utterances, and the entities
However, these features are taught by Weston:
for the given intent, prompting a large language model to generate a plurality of exemplar utterances that users could provide to perform the given intent, wherein the exemplar utterances comprise indications of entities (para. [0036]; the electronic device 14 may generate a list of training utterances for each intent in the list of intents. Generating the list of training utterances may occur automatically upon generating the list of intents. Each training utterance may be associated with an intent …, para. [0083]; the training utterances may be generated by a generative pretrained transformer (GPT) prompted to generate one or more utterances that include the intent …, para. [0091]-[0093]);
storing the given intent in a design-time format for a digital assistant (para. [0021]; para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0091]-[0093]; para. [0116])
creating a digital assistant definition with the exemplar utterances, and the entities (para. [0021]; para. [0056]; Each training utterance may be associated with an intent …, para. [0083]; para. [0091]-[0093]; para. [0116])
Gutierrez in view of Weston does not explicitly disclose loading a second large language model with documentation of APIs of the software application or for the given intent, prompting a third large language model
However, these features are taught by Mishra:
loading a second large language model with documentation of APIs of the software application (the system may include one or more machine learning model(s) other than one or more of the language models. …, col. 12, ln 50-59; The shortlister language model 740 processes the prompt data 715 to generate one or more API calls corresponding to request(s) that the corresponding APIs return a description of an action(s) that the APIs are configured to/will perform with respect to the user input and/or the current task…. the shortlister language model 740 may use the one or more exemplars included in the API descriptions (included in the prompt data 715) to determine the one or more input parameters for the API call.)
for the given intent, prompting a third large language model (the system may include one or more machine learning model(s) other than one or more of the language models. …, col. 12, ln 50-59; the personalized context component 565 (or the system 100) may include a personalized context prompt generation component (not illustrated), which may be configured to generate a prompt including the user input data 127 (or a representation of an intent of the user input) to be input to the LLM…., col. 17, ln 26-35)
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Weston with the method of Gutierrez in arriving at the missing features of Gutierrez, as well as to combine the teachings of Mishra with the method of Gutierrez in view of Weston in arriving at the missing features of Gutierrez in view of Weston, because such combination would have resulted in improving the accuracy or completeness of classifications with respect to input utterances (Weston, para. [0100]), as well as in enhancing the performance of a model and generating accurate responses (Mishra, col. 4, ln 9-31).
5. Claim 10 is rejected under 35 U.S.C. 103 as being unpatentable over Gutierrez in view of Mishra
Per claim 10, Gutierrez discloses the method of claim 1, wherein:
determining the invocation logic definition for the given intent comprises: converting an API definition into a natural language format (there may be an abstraction/translation of software and components (e.g., proprietary APIs and their data) into a universal language of data, capabilities … the various components are represented abstractly through Schemas which include Nouns 1920, Verbs 1930, Integrations …, para. [0332]; para. [0351]; para. [0473]; The API documentation generated could be in the format of an open-source API documentation standard such as OpenAPI in YAML, XML, or JSON …, para. [0498]-[0499]; certain methods (e.g., using large language models) can be used to automatically generate and/or partially generate SKL Schemas given other types of information about Integrations and data sources …, para. [0500]; para. [0505]; para. [0631]; para. [0684]); and
generating the invocation logic definition for the given intent (para. [0170]-[0177]; para. [0199]; para. [0238]; para. [0242]-[0246]; para. [0257]-[0263]; para. [0500]; para. [0737]).
Gutierrez does not explicitly disclose wherein generating the invocation logic definition for the given intent comprises applying a template to the natural language format.
However, this feature is taught by Mishra (col. 23, ln 21-56)
It would have been obvious to one of ordinary skill in the art before the effective filing of the invention to combine the teachings of Mishra with the method of Gutierrez in in arriving at the missing features of Gutierrez, because such combination would have resulted in enhancing the performance of a model and generating accurate responses (Mishra, col. 4, ln 9-31).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. See PTO 892 form.
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 OLUJIMI A ADESANYA whose telephone number is (571)270-3307. The examiner can normally be reached Monday-Friday 8:30-5:00pm.
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, Richemond Dorvil can be reached at 571-272-7602. 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.
/OLUJIMI A ADESANYA/Primary Examiner, Art Unit 2658
Prosecution Insights