DETAILED ACTION
This non-final office action is responsive to application 18/759,131 as submitted on June 28, 2024.
Claim status is currently pending and under examination for claims 1-20 of which independent claims are 1, 10 and 16.
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 Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Independent Claim 1
Step 2A Prong One: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, independent claim 1, under the broadest reasonable interpretation, recites the following limitations that are abstract ideas:
based on the input prompt, identifying intermediate response data used to generate an artificial intelligence (AI) response to the input prompt; (mental process)
The “identifying” step involves determining which intermediate response data to use based on an input prompt which amounts to no more than observations, evaluations, and judgments that can be performed in the human mind or with the use of a physical aid (e.g., pen and paper). The claim recites the step of identifying intermediate response data at a high degree of generality, thus the step is not required to have any specific level of complexity that would preclude the step from being mental processes. Therefore, the “identifying” step is considered to be mental processes, see MPEP § 2106.04(a)(2)(III).
Therefore, independent claim 1 recites a judicial exception.
Step 2A Prong Two: Does the claim recite additional elements that integrate the judicial exception into a practical application?
No, the judicial exception recited above is not integrated into a practical application. The claims recite the following additional elements, but these additional elements are not sufficient to integrate the judicial exception into a practical application:
a processor; (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea)
and computer storage memory having computer-executable instructions stored thereon that, when executed by the processor, configure the computing system to perform operations comprising: (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea)
obtaining an input prompt provided via a user interface; (MPEP § 2106.05(g) necessary data gathering and insignificant extra-solution activity to the judicial exception)
and providing the intermediate response data for presentation, via the user interface, in association with the AI response. (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
The “obtaining” step amounts to mere data gathering and is recited at a high level of generality, thus adding insignificant extra-solution activity to the judicial exception – see MPEP § 2106.05(g). Under MPEP § 2106.05(d), such additional elements have been found by the courts to not integrate a judicial exception into a practical application.
The “providing” step is recited at a high-level of generality such that the limitation amounts to no more than mere instructions to “apply” the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
The remaining additional elements are recited at a high-level of generality such that they amount to no more than mere instructions to “apply” an exception using a generic component. Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea, see MPEP § 2106.05(f).
Therefore, the above limitations do not integrate the judicial exception into a practical application.
Step 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. The claims do not include additional elements that are sufficient for the claims to amount to significantly more than the judicial exception.
In regards to the “obtaining” step, this step adds insignificant extra-solution activity. An extra-solution activity is a well-understood, routine and conventional (WURC) activity per MPEP § 2106.05(d)(II), “the courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity. i. Receiving or transmitting data over a network, e.g., using the Internet to gather data.” The “obtaining” step does not integrate the judicial exception into a practical application and does not amount to significantly more.
In regards to the “providing” step and the remaining additional elements, the limitations are recited so generically such that they amount to no more than mere instructions to “apply” the judicial exception on a computer using generic computer components. Mere instructions to apply a judicial exception cannot provide an inventive concept. See MPEP § 2106.05(f).
Therefore, independent claim 1 is not patent eligible.
Independent Claim 10
Step 2A Prong One: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, independent claim 10, under the broadest reasonable interpretation, recites the following limitations that are abstract ideas:
based on the input prompt, identifying, via the AI assistant manager, intermediate response data used to generate an AI response to the input prompt; (mental process)
The “identifying” step involves determining which intermediate response data to use based on an input prompt which amounts to no more than observations, evaluations, and judgments that can be performed in the human mind or with the use of a physical aid (e.g., pen and paper). The claim recites the step of identifying intermediate response data at a high degree of generality, thus the step is not required to have any specific level of complexity that would preclude the step from being mental processes. Therefore, the “identifying” step is considered to be mental processes, see MPEP § 2106.04(a)(2)(III).
Therefore, independent claim 10 recites a judicial exception.
Step 2A Prong Two: Does the claim recite additional elements that integrate the judicial exception into a practical application?
No, the judicial exception recited above is not integrated into a practical application. The claims recite the following additional elements, but these additional elements are not sufficient to integrate the judicial exception into a practical application:
obtaining, via an artificial intelligence (AI) assistant manager, an input prompt provided via a user interface; (MPEP § 2106.05(g) necessary data gathering and insignificant extra-solution activity to the judicial exception)
based on the input prompt, identifying, via the AI assistant manager, intermediate response data used to generate an AI response to the input prompt; (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
providing, via the AI assistant manager, the intermediate response data for presentation, via the user interface, in association with the AI response; (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
and obtaining user feedback in association with the presented intermediate response data to facilitate generation or presentation of subsequent intermediate response data. (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
The “obtaining … an input prompt” step amounts to mere data gathering and is recited at a high level of generality, thus adding insignificant extra-solution activity to the judicial exception – see MPEP § 2106.05(g). Under MPEP § 2106.05(d), such additional elements have been found by the courts to not integrate a judicial exception into a practical application.
The “identifying” step requires an “AI assistant manager” to identify intermediate response data. The AI assistant manager is used to apply the recited judicial exception without placing any limitation on how the AI assistant manager operates. The limitation amounts to mere instructions to “apply” the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
The “providing” step is recited at a high-level of generality such that the limitation amounts to no more than mere instructions to “apply” the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
The “obtaining user feedback” step is recited at a high-level of generality such that the limitation amounts to no more than mere instructions to “apply” the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
Therefore, the above limitations do not integrate the judicial exception into a practical application.
Step 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. The claims do not include additional elements that are sufficient for the claims to amount to significantly more than the judicial exception.
In regards to the “obtaining … an input prompt” step, this step adds insignificant extra-solution activity. An extra-solution activity is a well-understood, routine and conventional (WURC) activity per MPEP § 2106.05(d)(II), “the courts have recognized the following computer functions as well‐understood, routine, and conventional functions when they are claimed in a merely generic manner (e.g., at a high level of generality) or as insignificant extra-solution activity. i. Receiving or transmitting data over a network, e.g., using the Internet to gather data.” The step does not integrate the judicial exception into a practical application and does not amount to significantly more.
In regards to the “AI assistant manager” in the “identifying” step, the limitations are recited so generically such that they amount to no more than mere instructions to “apply” the judicial exception on a computer using generic computer components. Mere instructions to apply a judicial exception cannot provide an inventive concept. See MPEP § 2106.05(f).
In regards to the “providing” and “obtaining user feedback” steps, the limitations are recited so generically such that they amount to no more than mere instructions to “apply” the judicial exception on a computer using generic computer components. Mere instructions to apply a judicial exception cannot provide an inventive concept. See MPEP § 2106.05(f).
Therefore, independent claim 10 is not patent eligible.
Independent Claim 16
Step 2A Prong One: Does the claim recite an abstract idea, law of nature, or natural phenomenon?
Yes, independent claim 16, under the broadest reasonable interpretation, recites the following limitations that are abstract ideas:
identifying context data associated with an input prompt provided via a user interface; (mental process)
using the input prompt and at least a portion of the context data to generate a query and identify a data source for use in executing the query; (mental process)
The “identifying” step involves determining context data from an input prompt which amounts to no more than observations, evaluations, and judgments that can be performed in the human mind or with the use of a physical aid (e.g., pen and paper). The claim recites the step of identifying context data at a high degree of generality, thus the step is not required to have any specific level of complexity that would preclude the step from being mental processes. Therefore, the “identifying” step is considered to be mental processes, see MPEP § 2106.04(a)(2)(III).
The “using the input prompt …” step involves using an input prompt and context data to determine a query and a data source to retrieve data from which amounts to no more than observations, evaluations, and judgments that can be performed in the human mind or with the use of a physical aid (e.g., pen and paper). The claim recites the step at a high degree of generality, thus the step is not required to have any specific level of complexity that would preclude the step from being mental processes. Therefore, the step is considered to be mental processes, see MPEP § 2106.04(a)(2)(III).
Therefore, independent claim 16 recites a judicial exception.
Step 2A Prong Two: Does the claim recite additional elements that integrate the judicial exception into a practical application?
No, the judicial exception recited above is not integrated into a practical application. The claims recite the following additional elements, but these additional elements are not sufficient to integrate the judicial exception into a practical application:
One or more computer storage media having computer-executable instructions embodied thereon that, when executed by one or more processors, cause the one or more processors to perform a method, the method comprising: (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea)
based on execution of the query, obtaining a query result including data relevant to the query; (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
causing display, via the user interface, of at least one of the context data, the query, the data source, and the query result as an artificial intelligence (AI) response is being generated in association with the input prompt; (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
using one or more of the context data, the query, the data source, and the query result to generate, via a large language model, the AI response in association with the input prompt; (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
and causing display, via the user interface, of the AI response. (MPEP § 2106.05(f) mere instructions to implement an abstract idea on a computer, or generally links exception to a technological environment)
The “obtaining” step is recited at a high-level of generality such that the limitation amounts to no more than mere instructions to “apply” the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
The “display” steps are recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than attempts to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
The “using” step is recited at a high-level of generality such that the limitation amounts to no more than mere instructions to “apply” the judicial exception on a computer. It can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
The remaining additional elements are recited at a high-level of generality such that they amount to no more than mere instructions to “apply” an exception using a generic component. Adding the words “apply it” (or an equivalent) with the judicial exception, or mere instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea, see MPEP § 2106.05(f).
Therefore, the above limitations do not integrate the judicial exception into a practical application.
Step 2B: Does the claim recite additional elements that amount to significantly more than the judicial exception?
No. The claims do not include additional elements that are sufficient for the claims to amount to significantly more than the judicial exception.
In regards to the “obtaining”, “display”, and “using” steps, the limitations are recited so generically such that they amount to no more than mere instructions to “apply” the judicial exception on a computer using generic computer components. Mere instructions to apply a judicial exception cannot provide an inventive concept. See MPEP § 2106.05(f).
In regards to the remaining additional elements, the limitations are recited so generically such that they amount to no more than mere instructions to “apply” the judicial exception on a computer using generic computer components. Mere instructions to apply a judicial exception cannot provide an inventive concept. See MPEP § 2106.05(f).
Therefore, independent claim 16 is not patent eligible.
Dependent Claims 2-9, 11-15 and 17-20
The remaining dependent claims being rejected do not recite additional elements, whether considered individually or in combination, that are sufficient to integrate the judicial exception into a practical application or amount to significantly more than a judicial exception.
Claim limitation
Examiner analysis
2. The computing system of claim 1, wherein the intermediate response data comprises an indication of an intent associated with the input prompt, wherein the intent is identified based on analyzing the input prompt.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
3. The computing system of claim 1, wherein the intermediate response data comprises a skill associated with the input prompt, the skill corresponding with a function to assist with performing a task.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
4. The computing system of claim 1, wherein the intermediate response data comprises a query generated to obtain data relevant to the input prompt.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
5. The computing system of claim 1, wherein the intermediate response data comprises a data source for which to provide a query to obtain data relevant to the input prompt.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
6. The computing system of claim 1, wherein the intermediate response data comprises query results obtained in response to executing a query generated, based on the input prompt, to obtain data relevant to the input prompt.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
7. The computing system of claim 1, wherein the intermediate response data is provided as the AI response is being generated.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
8. The computing system of claim 1, wherein the intermediate response data is provided for concurrent presentation with the AI response via the user interface.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
9. The computing system of claim 1, wherein the intermediate response data is presented, via the user interface, based on receiving a user selection to view the intermediate response data.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
11. The computer-implemented method of claim 10, wherein the intermediate response data comprises an intent, a skill, a query, a data source, and/or a query result.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
12. The computer-implemented method of claim 10, wherein the user feedback comprises a presentation-related preference indicating a desired or undesired manner in which to present the intermediate response data.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
13. The computer-implemented method of claim 10, wherein the user feedback comprises a utilization-related preference indicating a desired or undesired manner in which to use the intermediate response data to generate the AI response.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
14. The computer-implemented method of claim 10 further comprising using the user feedback to train a large language model for subsequently generating a new AI response.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
15. The computer-implemented method of claim 10, wherein the intermediate response data is presented interleaved between the input prompt and the AI response.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
17. The media of claim 16, wherein the context data comprises an intent associated with the input prompt and/or a skill associated with the input prompt.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
18. The media of claim 16 further comprising obtaining user feedback associated with the display of the at least one of the context data, the query, the data source, and the query result.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
19. The media of claim 18 further comprising using the user feedback to modify the display of the at least one of the context data, the query, the data source, and the query result.
The step is recited at a high-level of generality such that the limitations amount to no more than mere instructions to “apply” the judicial exception on a computer. They can also be viewed as nothing more than an attempt to generally link the use of the judicial exception to the technological environment of computers, see MPEP § 2106.05(f).
20. The media of claim 18 further comprising using the user feedback to modify generation and/or presentation of a subsequently generated set of intermediate response data or a subsequently generated AI response.
This is a mental process akin to a human evaluation/judgment/observation.
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)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
The following is the reference relied upon in the rejections below:
Berglund (US 20240403341 A1)
Claims 1, 7-12 and 15-19 are rejected under 35 U.S.C. 102a(2) as being anticipated by Berglund.
Regarding Claim 1, Berglund teaches:
A computing system comprising ([0090] “the instruction(s) cause the computing system 500 to perform operations”):
a processor ([0090] processor);
and computer storage memory having computer-executable instructions stored thereon that, when executed by the processor, configure the computing system to perform operations comprising ([0090] “computer programs typically comprise one or more instructions (e.g., instructions 504, 508, 528) set at various times in various memory and storage devices in computing device(s). When read and executed by the processor 502, the instruction(s) cause the computing system 500 to perform operations”):
obtaining an input prompt provided via a user interface ([0032] “a user inputs a search query at the user interface 202”);
based on the input prompt, identifying intermediate response data used to generate an artificial intelligence (AI) response to the input prompt (The Examiner interprets “intermediate response data” according to its broadest reasonable interpretation (BRI) in view of the Applicant’s specification as encompassing search results. This interpretation is consistent with the illustrative descriptions in the Applicant’s specification at [0017, 0034], (see excerpts below).
Applicant’s written description at [0017]: “intermediate response data providing context or understanding as to how the response is generated is presented or displayed to a user. Various types of intermediate response data that may be presented include, without limitation, intermediate queries, raw query results, skills, intent, and/or the like.”
Applicant’s written description at [0034]: “types of intermediate response data that may be presented include context data (e.g., intent and/or skills), query data, source data, query results data, and/or the like.”
[0035] “the query service 112 invokes the search engine 114 to retrieve content items or portions of content items that are relevant to the search query and from which the LLM 140 should generate an answer.”
[0044] “Once content items related to the search query have been identified, the query service 112 can send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
[0046] "the query service 112 sends the LLM 140 a prompt that includes the search query and the text chunks matched to the search query, instructing the LLM 140 to generate an answer to the search query”
The query service invokes a search engine to retrieve content items that are relevant to a search query (‘input prompt’) to generate an answer (‘AI response’) to the search query. The retrieved content items are sent as a list of search results (query results) to a user interface. Since search results are used to generate an answer and are presented to a user during answer generation, the search results therefore help the user understand which search results are being used to generate the answer (and therefore the search results are intermediate response data).);
and providing the intermediate response data for presentation, via the user interface, in association with the AI response ([0044] “Once content items related to the search query have been identified, the query service 112 can send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
The query service sends (‘provides’) the search results (‘intermediate response data’) for display on a user interface while an answer (‘AI response’) is being generated. The search results are used to generate the answer (see [0035]), therefore the search results are provided in association with the answer.).
Regarding Claim 7, Berglund teaches:
The computing system of claim 1, wherein the intermediate response data is provided as the AI response is being generated ([0044] “the query service 112 can send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”).
Regarding Claim 8, Berglund teaches:
The computing system of claim 1, wherein the intermediate response data is provided for concurrent presentation with the AI response via the user interface ([0044] “send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
Berglund discloses Figure 3B (reproduced below) depicting a response (‘AI response’) and a list of search results (‘intermediate response data’).
PNG
media_image1.png
842
1488
media_image1.png
Greyscale
).
Regarding Claim 9, Berglund teaches:
The computing system of claim 1, wherein the intermediate response data is presented, via the user interface, based on receiving a user selection to view the intermediate response data (Berglund discloses Figure 3B (reproduced above) depicting a response and an option “Show me the list for 2021” on a user interface. When a user clicks on “Show me the list for 2021”, a list of search results (‘intermediate response data’) used to generate the response (answer) is displayed, therefore presenting intermediate response data based on receiving a user selection (click) to view the data.).
Regarding Claim 10, Berglund teaches:
A computer-implemented method comprising ([0090] “When read and executed by the processor 502, the instruction(s) cause the computing system 500 to perform operations”):
obtaining, via an artificial intelligence (AI) assistant manager, an input prompt provided via a user interface ([0032] “a user inputs a search query at the user interface 202, which can be displayed on a user device. The search query is received by the query service 112”
[0046] “the query service 112 sends the LLM 140 a prompt that includes the search query”
A query service receives a user search query (‘input prompt’) and sends it to an LLM, therefore the query service handles and routes data for the LLM and is therefore an AI assistant manager.);
based on the input prompt, identifying, via the AI assistant manager, intermediate response data used to generate an AI response to the input prompt (The Examiner interprets “intermediate response data” according to its BRI in view of the Applicant’s specification as encompassing search results. This interpretation is consistent with the illustrative descriptions in the Applicant’s specification at [0017, 0034].
[0035] “the query service 112 invokes the search engine 114 to retrieve content items or portions of content items that are relevant to the search query and from which the LLM 140 should generate an answer.”
[0044] “Once content items related to the search query have been identified, the query service 112 can send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
[0046] "the query service 112 sends the LLM 140 a prompt that includes the search query and the text chunks matched to the search query, instructing the LLM 140 to generate an answer to the search query”
The query service (‘AI assistant manager’) invokes a search engine to retrieve content items that are relevant to a search query (‘input prompt’) to generate an answer (‘AI response’) to the search query. The retrieved content items are sent as a list of search results (query results) to a user interface. Since search results are used to generate an answer and are presented to a user during answer generation, the search results therefore help the user understand which search results are being used to generate the answer (and therefore the search results are intermediate response data).);
providing, via the AI assistant manager, the intermediate response data for presentation, via the user interface, in association with the AI response ([0044] “Once content items related to the search query have been identified, the query service 112 can send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
The query service (‘AI assistant manager’) sends (‘provides’) the search results (‘intermediate response data’) for display on a user interface while an answer (‘AI response’) is being generated. The search results are used to generate the answer (see [0035]), therefore the search results are provided in association with the answer.);
and obtaining user feedback in association with the presented intermediate response data to facilitate generation or presentation of subsequent intermediate response data ([0053] “The user can provide feedback on the answer, for example to indicate that the user liked/disliked the answer or that the answer was helpful/unhelpful. The feedback can additionally or alternatively include more granular signals, such as the answer not matching what the user was looking for, the answer being confusing, etc. The feedback received from the user can be used by the content management system to improve the operations of the content management system or the LLMs described herein.”
[0021] “the content management system 110 can include a query service 112 and a search engine 114. Using the query service 112 and search engine 114, the content management system 110 facilitates user searches of the content repository 150 that return answers to the users' queries. … These answers can be provided instead of or in addition to search results that identify items in the content repository semantically related to the search query.”
Search results (‘presented intermediate response data’) can be provided as an answer. The user can provide feedback on the answer (search results) to improve operations of a content management system (search engine) that retrieves search results. Therefore, when a search engine retrieves content items that are sent as a list of search results for display (see [0044]), the improvements to the content management system would improve how the search engine retrieves the content items (provide more relevant search results) and therefore facilitating generation of more relevant, subsequent lists of search results (‘subsequent intermediate response data’) for display based on user feedback.).
Regarding Claim 11, Berglund teaches:
The computer-implemented method of claim 10, wherein the intermediate response data comprises an intent, a skill, a query, a data source, and/or a query result (Berglund discloses search results (‘query result’) are displayed, “the query service 112 can send, at 224, a list of the search results to the user interface for display” [0044].).
Regarding Claim 12, Berglund teaches:
The computer-implemented method of claim 10, wherein the user feedback comprises a presentation-related preference indicating a desired or undesired manner in which to present the intermediate response data (Berglund discloses Figure 3B (reproduced above) depicting a response and an option “Show me the list for 2021” on a user interface. When a user clicks on “Show me the list for 2021”, a list of search results (‘intermediate response data’) used to generate the response (answer) is displayed, otherwise the search results are hidden. By clicking on “Show me the list for 2021”, user feedback in the form of a click is received. The click (user feedback) indicates that a user desires (prefers) to have all the search results displayed to view, instead of being hidden.).
Regarding Claim 15, Berglund teaches:
The computer-implemented method of claim 10, wherein the intermediate response data is presented interleaved between the input prompt and the AI response ([0044] “send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
Berglund discloses Figure 3B (reproduced above) depicting a chat on a user interface. A search query “What are the key product launches this quarter?” and a response (‘AI response’) is displayed. A list of search results (‘intermediate response data’) is sent to the user interface while an answer is being generated, therefore the order the user is presented information is the search query (‘input prompt’), then the search results (‘intermediate response data’) and then the response (answer). Therefore, intermediate response data is presented interleaved between the input prompt (search query) and the AI response.).
Regarding Claim 16, Berglund teaches:
One or more computer storage media having computer-executable instructions embodied thereon that, when executed by one or more processors, cause the one or more processors to perform a method, the method comprising ([0090] “computer programs typically comprise one or more instructions (e.g., instructions 504, 508, 528) set at various times in various memory and storage devices in computing device(s). When read and executed by the processor 502, the instruction(s) cause the computing system 500 to perform operations”):
identifying context data associated with an input prompt provided via a user interface ([0040] “relevancy criterion is based on context of a search query. In some implementations, the query service 112 also receives information about a context of the query and passes the context information to the search engine 114 to filter or refine the set of content items that are searched or matched to a search query. The context information can include, for example, an identity of a user who submitted the query, a history of queries recently submitted by the user, or a context in which the query was submitted”
[0032] “a user inputs a search query at the user interface 202”);
using the input prompt and at least a portion of the context data to generate a query and identify a data source for use in executing the query ([0040] “the query service 112 can send the context information to the LLM 140 as part of a prompt to generate an answer to the search query.”
[0046] “the query service 112 sends the LLM 140 a prompt that includes the search query”
[0035-0036] “the query service 112 invokes the search engine 114 to retrieve content items or portions of content items that are relevant to the search query and from which the LLM 140 should generate an answer … search engine 114 to retrieve the specified content item from the repository 150”
A query service uses a search query (‘input prompt’) and context information (‘context data’) to generate a prompt (‘query’) to send to an LLM. The query service invokes a search engine to retrieve content items that are relevant to the search query from repository 150 (‘data source’), therefore identifying a data source for use in executing the generated prompt.);
based on execution of the query, obtaining a query result including data relevant to the query ([0044] “Once content items related to the search query have been identified, the query service 112 can send, at 224, a list of the search results”
[0035] “retrieve content items or portions of content items that are relevant to the search query”);
causing display, via the user interface, of at least one of the context data, the query, the data source, and the query result as an artificial intelligence (AI) response is being generated in association with the input prompt ([0044] “Once content items related to the search query have been identified, the query service 112 can send, at 224, a list of the search results to the user interface for display while an answer to the user's search query is generated.”
A list of search results (‘query result’) is displayed on a user interface while an answer (‘AI response’) is being generated for a user’s search query (‘input prompt’), therefore an answer is associated with a search query (‘input prompt’).);
using one or more of the context data, the query, the data source, and the query result to generate, via a large language model, the AI response in association with the input prompt ([0040] “send the context information to the LLM 140 as part of a prompt to generate an answer to the search query.”
[0046] “sends the LLM 140 a prompt that includes the search query and the text chunks matched to the search query, instructing the LLM 140 to generate an answer to the search query”
[0035-0036] “invokes the search engine 114 to retrieve content items or portions of content items that are relevant to the search query and from which the LLM 140 should generate an answer … search engine 114 to retrieve the specified content item from the repository 150”
A prompt (‘the query’) is generated from context information (‘context data’) and a user’s search query (‘input prompt’) to instruct an LLM (‘large language model’) to generate an answer (‘AI response’) to the search query. The search engine retrieves content items (‘query result’) from a repository (‘data source’) that are used to generate the answer.);
and causing display, via the user interface, of the AI response ([0053] “Once the query service 112 has received an answer from an LLM and optionally validated the answer, the query service 112 provides the answer for output to the user, at operation 230 in FIG. 2B. For example, the query service 112 returns text-based answers for display within the user interface 202.”).
Regarding Claim 17, Berglund teaches:
The media of claim 16, wherein the context data comprises an intent associated with the input prompt and/or a skill associated with the input prompt ([0047] “The prompt to the LLM 14 can further include context information of the query. For example, the query service 112 retrieves contextual information associated with the user who submitted the search query or the context in which the search query was generated, … the query service 112 retrieves a history of content accessed, generated, or shared by the user and identifies activity that relates to the content used as the knowledge base for the answer. If, for example, a user recently attached a document entitled “Strategic Framework” to a draft email to a sales prospect and subsequently submitted a search query for “Explain how the strategic framework benefits companies in the life sciences industry,” the query service 112 can predict that the user is requesting an introductory sales pitch that is suitable for an email.”
User activity (attaching a document to an email) and a submitted search query (‘input prompt’) is used to predict what a user is requesting with the submitted search query (the user’s intent), therefore context information of a submitted search query comprises intent associated with the query.).
Regarding Claim 18, Berglund teaches:
The media of claim 16 further comprising obtaining user feedback associated with the display of the at least one of the context data, the query, the data source, and the query result (Berglund discloses Figure 3B (reproduced above) depicting a response and an option “Show me the list for 2021” on a user interface. When a user clicks on “Show me the list for 2021”, a list of search results (‘query result’) is displayed, otherwise the search results are hidden. By clicking on “Show me the list for 2021”, user feedback in the form of a click is received. The click (user feedback) indicates that a user desires to have all the search results displayed to view, instead of being hidden.).
Regarding Claim 19, Berglund teaches:
The media of claim 18 further comprising using the user feedback to modify the display of the at least one of the context data, the query, the data source, and the query result (Berglund discloses Figure 3B depicting a response and an option “Show me the list for 2021” on a user interface. By clicking on “Show me the list for 2021”, user feedback in the form of a click is received. The click (user feedback) indicates that a user desires to have all the search results (‘query result’) displayed to view, instead of being hidden (therefore using user feedback (clicks) to modify the display of search results).).
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 following are the references relied upon in the rejections below:
Weidele, Daniel Karl I., et al. "Empirical evidence on conversational control of GUI in semantic automation." Proceedings of the 29th International Conference on Intelligent User Interfaces. 2024.
Claims 2-3 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Berglund / Weidele.
Regarding Claim 2, Berglund teaches: The computing system of claim 1, however, Berglund does not teach intermediate response data comprising an indication of an intent associated with an input prompt, which is taught by Weidele:
wherein the intermediate response data comprises an indication of an intent associated with the input prompt (The Examiner interprets “intermediate response data” according to its BRI in view of the Applicant’s specification as encompassing a chat response. This interpretation is consistent with the illustrative descriptions in the Applicant’s specification at [0017, 0034].
(P. 870, Figure 1 Caption) “User conversationally asks to add geographic data to a selected table (yellow dot). Chatbot recognizes the intent (table join-ability search), extracts entities (table id from UI context, ”geography” as search keyword) and generates a chat response. State of the UI adjusts accordingly to reflect chatbot’s decisions”
(P. 875, Sec. 4.1, ¶5) “The agent recognizes this intent as a search for join-able tables [R3] and pulls up the corresponding table expansion view injecting the contextual table along with an extracted keyword (ref. figure 1 on the front page of this paper)”
Weidele discloses Figure 1 on P. 870 (reproduced below) depicting a chat between a user and a chatbot. The user asks to add geographic data to a table (‘input prompt’). The chatbot replies with a response describing that it will search for geographical information in a lakehouse, therefore repeating the user’s intent for adding geographic data and describing how it will generate its final chat response (a joint table) using the lakehouse. Therefore, the chat response describing the user’s intent and the steps the chatbot will take is ‘intermediate response data’.
PNG
media_image2.png
560
760
media_image2.png
Greyscale
),
wherein the intent is identified based on analyzing the input prompt ((P. 870, Figure 1 Caption) “User conversationally asks to add geographic data to a selected table (yellow dot). Chatbot recognizes the intent (table join-ability search)”).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the technique disclosed by Weidele to include intent in intermediate responses to users. By including intent in intermediate responses to users, a model can show a user its interpretation of their query before a final response is generated, thereby allowing a user to correct wrong interpretations and understand why a model generated the final response.
Regarding Claim 3, Berglund teaches The computing system of claim 1, however Berglund does not teach intermediate response data comprising a skill associated with an input prompt, which is taught by Weidele:
wherein the intermediate response data comprises a skill associated with the input prompt, the skill corresponding with a function to assist with performing a task (The Examiner interprets “intermediate response data” according to its BRI in view of the Applicant’s specification as encompassing a chat response. This interpretation is consistent with the illustrative descriptions in the Applicant’s specification at [0017, 0034].
(P. 870, Figure 1 Caption) “User conversationally asks to add geographic data to a selected table (yellow dot). Chatbot recognizes the intent (table join-ability search), extracts entities (table id from UI context, ”geography” as search keyword) and generates a chat response. State of the UI adjusts accordingly to reflect chatbot’s decisions”
(P. 875, Sec. 4.1, ¶5) “The agent recognizes this intent as a search for join-able tables [R3] and pulls up the corresponding table expansion view injecting the contextual table along with an extracted keyword (ref. figure 1 on the front page of this paper)”
Weidele discloses Figure 1 on P. 870 (reproduced above) depicting a chat between a user and a chatbot. The user asks to add geographic data to a table (‘input prompt’). The chatbot replies with a response describing that it will see if there are compatible tables with geographical information in the lakehouse, therefore describing it will perform a search (‘function to assist with performing a task’), and therefore a search is a skill used to assist with performing a task of generating a joint table. The chat response describes it will perform a search (skill) to generate its final response (a joint table), therefore the response is ‘intermediate response data’.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the technique disclosed by Weidele to describe an action that will be performed in intermediate responses to users. By describing an action that will be performed in intermediate responses to users, users can determine if a model correctly interpreted their query if the model initiates an appropriate action, thereby allowing users to correct the model before a final response is generated and understand which actions were performed to generate the final response.
Regarding Claim 20, Berglund teaches: The media of claim 18, however, Berglund does not teach using user feedback to modify generation and/or presentation or subsequently generated intermediate response data or a generated AI response, which is taught by Weidele:
further comprising using the user feedback to modify generation and/or presentation of a subsequently generated set of intermediate response data or a subsequently generated AI response ((P. 870, Abstract) “This research explores integration of a Large Language Model (LLM) fine-tuned to conversationally control the user interface (UI) for a Semantic Automation Layer (SAL). … We augment the UI with our conversational engine and propose In-situ Prompt Engineering and learn from Human Feedback to smoothen the interaction and manipulation of UI through natural language commands.”
(P. 977, Sec. 4.2, Last Paragraph) “To collect feedback in production, hovering over chat responses allows users to engage by up- or down-vote. … For Data Engineers, the UI also features a view to provide feedback on the Semantic Enrichments in which generated captions, descriptions, tags or concept mappings can be corrected for a given table”
(P. 875, Sec. 4.1, ¶2) “The user formulates their utterance in the chat, which agent recognizes as a search. Agent then pulls up the search view in the UI and injects appropriate search keyword before terminating the request with a chat response [R2]. In the search view, we retrieve data tables from the lakehouse and present the results ranked by semantic relevance to the keyword. The otherwise raw technical table names are replaced with clean, AI-generated table captions”
A user enters an utterance into a chat that an agent recognizes as a search request (query). The agent returns data tables (query results) and displays them with AI-generated captions in a UI. Users can provide feedback to correct the AI-generated captions, which an LLM that controls the UI can learn from. Therefore, the LLM learns from user feedback to modify generation and presentation of subsequently generated AI responses (AI-generated captions).).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the user feedback technique disclosed by Weidele to use user feedback to modify the generation and display of AI responses. By using user feedback to modify the generation and display of AI responses, an LLM can learn that its generated responses are not correct from feedback, thereby enabling the LLM to generate and display responses that are more accurate when processing similar prompts.
The following are the references relied upon in the rejections below:
Yu, Lei, and Abir Ray. "An llm maturity model for reliable and transparent text-to-query." arXiv preprint arXiv:2402.14855 (2024).
Claims 4-6 are rejected under 35 U.S.C. 103 as being unpatentable over Berglund / Yu.
Regarding Claim 4, Berglund teaches: The computing system of claim 1, however, Berglund does not teach intermediate response data comprises a generated query, which is taught by Yu:
wherein the intermediate response data comprises a query generated to obtain data relevant to the input prompt ((P. 2, Sec. 2, ¶2) “Imagine a scenario where a law enforcement agent obtains a phone number from a source and seeks to: 1) determine if the number is linked to any suspects in a database of individuals with criminal records; … the agent can effortlessly inquire about complex data without prior knowledge of the underlying data sources or how to formulate database queries. Through a simple chat interface, the agent can pose a series of questions and immediately receive precise results in desired formats … QueryIQ ensures traceability and transparency, allowing the gent to trace intermediate steps leading to answers and examine the database queries generated by the underlying LLM to retrieve relevant data elements. This enables agents to iteratively prompt the system with further inquiries, facilitating deeper investigation.”
Yu discloses Figure 1 on P. 2 (reproduced below) depicting a chat interface of a conversation between a law enforcement agent and QueryIQ system. The agent enters a question (‘input prompt’) and receives intermediate steps (generated SQL query) that lead to an answer (list of suspects). The SQL query is generated based on the agent’s question to obtain a list of suspects, therefore the SQL query (intermediate steps) is ‘intermediate response data’.
PNG
media_image3.png
898
1378
media_image3.png
Greyscale
).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the intermediate steps disclosed by Yu to output intermediate steps containing queries. By outputting intermediate steps containing queries, users can understand which queries were used to retrieve results used in generating an answer, thereby increasing traceability and transparency.
Regarding Claim 5, Berglund teaches: The computing system of claim 1, however, Berglund does not teach intermediate response data comprising a data source, which is taught by Yu:
wherein the intermediate response data comprises a data source for which to provide a query to obtain data relevant to the input prompt ((P. 2, Sec. 2, ¶2) “Through a simple chat interface, the agent can pose a series of questions and immediately receive precise results in desired formats … QueryIQ ensures traceability and transparency, allowing the gent to trace intermediate steps leading to answers and examine the database queries generated by the underlying LLM to retrieve relevant data elements”
Yu discloses Figure 1 on P. 2 (reproduced above) depicting a chat interface of a conversation between a law enforcement agent and QueryIQ system. The agent enters a question (‘input prompt’) and receives intermediate steps (generated SQL query) that lead to an answer (list of suspects). The SQL query is generated based on the agent’s question to obtain a list of suspects, therefore the SQL query (intermediate steps) is ‘intermediate response data’. The SQL query contains databases and tables (‘data sources’) for which the query retrieves data from, therefore the ‘intermediate response data’ comprises data sources.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the intermediate steps disclosed by Yu to output intermediate steps containing data sources. By outputting intermediate steps containing data sources, users can determine which databases or tables were used to retrieve results to generate an answer, thereby increasing traceability and transparency.
Regarding Claim 6, Berglund teaches: The computing system of claim 1, however Berglund does not teach intermediate response data comprising query results obtained in response to executing a generated query, which is taught by Yu:
wherein the intermediate response data comprises query results obtained in response to executing a query generated, based on the input prompt, to obtain data relevant to the input prompt ((P. 2, Sec. 2, ¶2) “Through a simple chat interface, the agent can pose a series of questions and immediately receive precise results in desired formats … allowing the gent to trace intermediate steps leading to answers and examine the database queries generated by the underlying LLM to retrieve relevant data elements”
(P. 8, Exhibit 2 Caption) “The LLM retrieves intermediate result of all tolls data related to the given phone number before producing the top 10 most contacted numbers for the suspect.”
Yu discloses Exhibit 2 on P. 8 (reproduced below) depicting a chat interface of a conversation between a law enforcement agent and QueryIQ system. The agent enters a question (‘input prompt’) and receives intermediate results (all tolls data) that lead to an answer (top 10 most contacted numbers). The intermediate results (‘intermediate response data’) are obtained by executing an SQL query generated based on the agent’s question. Therefore, the intermediate results are query results obtained in response to executing an SQL query.
PNG
media_image4.png
740
1207
media_image4.png
Greyscale
).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the intermediate results disclosed by Yu to output intermediate results containing all results returned from executing a query. By outputting intermediate results containing all results returned from executing a query, a user can compare the intermediate results with the results from a final answer to identify which results were included or excluded in the answer, thereby helping a user understand which results were filtered and why.
The following are the references relied upon in the rejections below:
Ruan, Jingqing, et al. "Tptu: Large language model-based ai agents for task planning and tool usage." arXiv preprint arXiv:2308.03427v1 (2023).
Claims 13-14 are rejected under 35 U.S.C. 103 as being unpatentable over Berglund / Raun.
Regarding Claim 13, Berglund teaches: The computer-implemented method of claim 10, however Berglund does not teach user feedback comprising a utilization-related preference, which is taught by Raun:
wherein the user feedback comprises a utilization-related preference indicating a desired or undesired manner in which to use the intermediate response data to generate the AI response ((P. 4, Sec. 2.1, Bullet Point 5) “Intermediate Output. This represents the output generated by the LLM-based AI Agent after it processes the task instructions and prompts, and interacts with the toolset. There are three typical intermediate outputs: (1) the high-level plans to fulfill the original user instruction, (2) selected and created tools to fulfill each subtask in the plans, and (3) the results or errors produced after tool execution. The output can be reviewed and refined, either by the AI Agent itself or with human oversight, to ensure it is accurate and meets the requirements of the task instruction.”
(P. 4, Sec. 2.1, Bullet Point 6) “Final Answer. This is the output that the AI Agent summarizes and provides to the user after all processing (including task planning, tool usage, and possibly error feedback) has been completed.”
A final answer (‘AI response’) to an original user instruction is generated by processing intermediate outputs (task planning, tool usage and error feedback). Intermediate outputs (‘intermediate response data’) can be reviewed and refined by a human user, therefore when a user refines the intermediate outputs, the user changes an intermediate output according to their needs and corrects the output, therefore user refinement is user feedback that comprises a utilization-related preference indicating a desired manner (refining intermediate outputs according to user needs) in which to use the intermediate outputs to generate a final answer.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the intermediate output refinement technique disclosed by Raun to refine intermediate outputs. By refining intermediate outputs, users can modify or correct intermediate outputs according to their needs, thereby ensuring that the intermediate outputs used to generate a final answer are as accurate and useful as possible.
Regarding Claim 14, Berglund teaches: The computer-implemented method of claim 10, however, Berglund does not teach using user feedback to train a large language model to generate a new AI response which is taught by Raun:
further comprising using the user feedback to train a large language model for subsequently generating a new AI response (P. 4, Sec. 2.1, Bullet Point 5) “Intermediate Output. This represents the output generated by the LLM-based AI Agent after it processes the task instructions and prompts, and interacts with the toolset. There are three typical intermediate outputs: (1) the high-level plans to fulfill the original user instruction, (2) selected and created tools to fulfill each subtask in the plans, and (3) the results or errors produced after tool execution. The output can be reviewed and refined, either by the AI Agent itself or with human oversight, to ensure it is accurate and meets the requirements of the task instruction.”
(P. 4, Sec. 2.1, Bullet Point 6) “Final Answer. This is the output that the AI Agent summarizes and provides to the user after all processing (including task planning, tool usage, and possibly error feedback) has been completed.”
(P. 4, Sec. 2.2, ¶1) “AI Agents should be capable of learning from feedback, including correct results and exception errors. They should incorporate memory, such as logging or chat history, and reflection to adapt their plans or decisions. This allows the agents to continuously improve their performance and efficiency in task execution.”
A final answer (‘AI response’) to an original user instruction is generated by processing intermediate outputs (plan of subtasks). Intermediate outputs can be reviewed and refined by a human user, therefore user refinement is user feedback. An LLM-based AI agent can learn from feedback (human review) to improve how it plans task execution (changes how it generates a plan of subtasks used to generate a final answer), therefore using user feedback to train an LLM to generate new AI responses.).
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify the method of Berglund with the technique disclosed by Raun to train an LLM with user feedback. By training an LLM with user feedback, the LLM can learn from the feedback which intermediate outputs are erroneous, thereby improving how the LLM generates intermediate outputs and preventing erroneous intermediate outputs from being used in a final answer.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Gao et al. ("Self-explanation prompting improves dialogue understanding in large language models.") teaches improving the understanding of a dialogue between a user and LLM by having the LLM explain a user’s prompt before executing a task.
Ho et al. (US 20250291863 A1) teaches outputting intermediate search results to a user before the user responds to a query prompt that refines the user’s initial search query.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PEDRO J MORALES whose telephone number is (571)272-6106. The examiner can normally be reached 8:30 AM - 6:00 PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, MIRANDA M HUANG can be reached at (571)270-7092. 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.
/PEDRO J MORALES/Examiner, Art Unit 2124
/MIRANDA M HUANG/Supervisory Patent Examiner, Art Unit 2124