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 § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 21 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 21 recites “in particular numerical data or numerical values of.” One skilled in the art could not determine the scope of this claim because there is nothing after the proposition “of.” This renders the claim vague and indefinite.
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 1-6, 8, and 10-20 are rejected under 35 USC 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 recites
1. A system for processing input data, the system comprising: one or more processors; an analytical artificial intelligence module that is, when executed by at least one of the one or more processors, configured to
determine, metadata for the input data and weights of the metadata
a prompting module that is, when executed by at least one of the one or more processors, configured to
determine, based at least on a part of the metadata, and the weights of the metadata, a prompt;
and a generative artificial intelligence module that is, when executed by at least one of the one or more processors, configured to
determine, based on the prompt, output data.
Examiner finds that the emphasized portions of claim 1 above recite an abstract idea—namely, mental processes. See MPEP 2106.04(a)(2)(III):
Accordingly, the ‘mental processes’ abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, judgments, and opinions
When read as a whole, the recited limitations are directed to using mental steps to observe, evaluate, and make judgements about electronic data.
Taking each element individually, Examiner provides the following analysis:
Bolded Abstract Idea Claim Elements
Examiner analysis of bolded abstract idea elements considered individually
determine, metadata for the input data and weights of the metadata
This element merely requires observation and evaluation of the input data and an evaluation/ judgment as to how to determine the metadata and the weights.
determine, based at least on a part of the metadata and the weights of the metadata, a prompt;
This element merely requires observation and evaluation of the metadata and the weights and an evaluation/ judgment as to how to determine the prompt.
determine, based on the prompt, output data.
This element merely requires observation and evaluation of the prompt and an evaluation/ judgment as to how to determine the output data.
Turning to the additional elements, Examiner provides the following analysis:
Italicized Additional elements
Examiner analysis of italicized additional elements and whether they integrate the exception and whether they recite an inventive concept.
Relevant MPEP sections
1. A system for processing input data, the system comprising: one or more processors; an analytical artificial intelligence module that is, when executed by at least one of the one or more processors, configured to
These elements recite mere instructions to apply the exception. Thus, they do not integrate the exception and do not recite an inventive concept.
These elements generally link the abstract idea to a computing environment. Thus, they are field of use limitations that do not integrate the exception and do not recite an inventive concept.
2106.05(f),(h)
a prompting module that is, when executed by at least one of the one or more processors, configured to
These elements recite mere instructions to apply the exception. Thus, they do not integrate the exception and do not recite an inventive concept.
These elements generally link the abstract idea to a computing environment. Thus, they are field of use limitations that do not integrate the exception and do not recite an inventive concept.
2106.05(f),(h)
and a generative artificial intelligence module that is, when executed by at least one of the one or more processors, configured to
These elements recite mere instructions to apply the exception. Thus, they do not integrate the exception and do not recite an inventive concept.
These elements generally link the abstract idea to a generative artificial intelligence environment. Thus, they are field of use limitations that do not integrate the exception and do not recite an inventive concept.
2106.05(f),(h)
The additional elements above “[a]dd nothing … that is not already present when the steps are considered separately’”. MPEP 2106.05 (I)(B)(quoting Alice).
As such, when the claim elements are considered as a whole and individually, claim 1 recites an abstract idea without significantly more.
Claim 19 recites
19. A computer-implemented method for improving human-machine interaction, the method comprising:
feeding input data to an analytical artificial intelligence module for
determining metadata for the input data and weights of the metadata
feeding at least a part of the metadata to a prompting module for
determining a prompt for the input data, the prompt including metadata which are at least one of ranked in accordance with the weights of the metadata and selected based on the weights of the metadata
feeding the prompt to a generative artificial intelligence module for
determining output data for the input data.
Examiner finds that the emphasized portions of claim 19 above recite an abstract idea—namely, mental processes and mathematical concepts. See MPEP 2106.04(a)(2)(III):
Accordingly, the ‘mental processes’ abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, judgments, and opinions
When read as a whole, the recited limitations are directed to using mental steps to observe, evaluate, and make judgements about electronic data and mathematical calculations.
Taking each element individually, Examiner provides the following analysis:
Bolded Abstract Idea Claim Elements
Examiner analysis of bolded abstract idea elements considered individually
determining metadata for the input data and weights of the metadata
This element merely requires observation and evaluation of the input data and an evaluation/ judgment as to how to determine the metadata and the weights.
determining a prompt for the input data, the prompt including metadata which are at least one of ranked in accordance with the weights of the metadata and selected based on the weights of the metadata
This element merely requires observation and evaluation of the metadata and the weights and an evaluation/ judgment as to how to determine the prompt according to the metadata and the weights.
determining output data for the input data.
This element merely requires observation and evaluation of the prompt and an evaluation/ judgment as to how to determine the output data.
Turning to the additional elements, Examiner provides the following analysis:
Italicized Additional elements
Examiner analysis of italicized additional elements and whether they integrate the exception and whether they recite an inventive concept.
Relevant MPEP sections
19. A computer-implemented method for improving human-machine interaction, the method comprising:
These elements recite mere instructions to apply the exception. Thus, they do not integrate the exception and do not recite an inventive concept.
These elements generally link the abstract idea to a computing environment. Thus, they are field of use limitations that do not integrate the exception and do not recite an inventive concept.
2106.05(f),(h)
feeding input data to an analytical artificial intelligence module for
This element generally links the abstract idea to the field of use of analytical artificial intelligence. Thus, it is field of use limitation that does not integrate the exception and does not recite an inventive concept.
2106.05 (h)
feeding at least a part of the metadata to a prompting module for
This elements generally link the abstract idea to a computing environment. Thus, they are field of use limitations that do not integrate the exception and do not recite an inventive concept.
2106.05((h)
feeding the prompt to a generative artificial intelligence module for
This element generally links the abstract idea to the field of use of generative artificial intelligence. Thus, it is field of use limitation that does not integrate the exception and does not recite an inventive concept.
2106.05 (h)
The additional elements above “[a]dd nothing … that is not already present when the steps are considered separately’”. MPEP 2106.05 (I)(B)(quoting Alice).
The above analysis applies to claim 20 also. The limitation “A computer program product or a non-volatile computer-readable medium comprising instructions which, when executed by one or more processors of a system, cause the system to carry out the following processes” recites mere instructions to apply the exception and does not integrate or recite an inventive concept.
As such, when the claim elements are considered as a whole and individually, claim 19 and claim 20 recite an abstract idea without significantly more.
Dependent claims below are rejected under 35 USC 101 for the reasons indicated below.
Claim
Bold = abstract idea elements
Italics = additional elements
Analysis
MPEP
2. The system of claim 1, wherein the analytical artificial intelligence module comprises:
at least one preprocessing module that is, when executed by at least one of the one or more processors, configured to at least one of:
This element recite mere instructions to apply the exception and generally links the abstract idea to a computing environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
linguistically process the input data;
This element merely requires evaluation of the input data.
2106.04(a)(2)(III)
an analytical artificial intelligence metadata generating module that is, when executed by at least one of the one or more processors, configured to at least one of:
This element recite mere instructions to apply the exception and generally links the abstract idea to an analytical artificial intelligence environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
determining the respective metadata in accordance with the respective metadata request;
This element merely requires observation and evaluation as to what the “respective metadata” is according to the request.
2106.04(a)(2)(III)
3. (Currently Amended) The system of claim 1, wherein the metadata comprise at least one of: a semantic metadata related with the input data, numerical metadata, quantitative metadata, qualitative metadata, and content metadata for a content of the input data, a category of the content or a topic of the content, a concept related with the content and/or the category, at least one keyword, an entity related with the content, the concept and/or the category, and/or at least one relevant phrase related with the content, the concept and/or the category.
This element merely recites the data being evaluated.
2106.04(a)(2)(III)
4. The system of claim 1, wherein the system is a hybrid artificial intelligence system, wherein the system is a heterogeneous artificial intelligence system,
This element recite mere instructions to apply the exception and generally links the abstract idea to an hybrid/heterogeneous intelligence environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
wherein the input data comprise input textual data, wherein the output data comprise output textual data,
This element merely recites the data being evaluated.
2106.04(a)(2)(III)
wherein the generative artificial intelligence module is based on machine learning, implemented in a container, configured to generate the output text data, and/or comprises a Large Language Model,
This element recite mere instructions to apply the exception and generally links the abstract idea to an machine learning/generative AI and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
and wherein the analytical artificial intelligence module is based on machine learning, implemented in a container, and/or comprises at least one of:
This element recite mere instructions to apply the exception and generally links the abstract idea to an analytical AI and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
an intent analysis capability
Analyzing intent merely requires human observation, evaluation, and judgment.
2106.04(a)(2)(III)
5. The system of claim 1, wherein the analytical artificial intelligence module is, when executed by at least one of the one or more processors, configured to
This element recite mere instructions to apply the exception and generally links the abstract idea to analytical artificial intelligence and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
determine, based on the metadata for the input data, a request for at least one of: a user feedback and further input data.
This element merely requires observation and evaluation of the metadata and an evaluation/judgment as to what constitutes user feedback or further input data.
2106.04(a)(2)(III)
6. The system of claim 5, wherein the analytical artificial intelligence module is, when executed by at least one of the one or more processors, configured to
This element recite mere instructions to apply the exception and generally links the abstract idea to analytical artificial intelligence and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
update the metadata in accordance with a received user feedback and/or based on received further input data.
This element recites insignificant extra solution activity in the form of selecting a particular data source or type of data to be manipulated. It does not integrate the exception.
This element recites a well-understood, routine, and conventional (WURC) computer function (storing and receiving information in memory and/or electronic recordkeeping) and thus fails to recite an inventive concept.
2106.05(g)
2106.05(d) (II)
8. The system of claim 1, wherein the prompting module is, when executed by at least one of the one or more processors, configured to at least one of:
See claim 1 above.
selecting based on the weights of the metadata, metadata to be included into the prompt;
This element merely requires evaluation of the weights and an evaluation/judgment as to which metadata to select to include in the prompt.
2106.04(a)(2)(III)
10. (Currently Amended) The system of claim 1, wherein the input data are received via an interface from a user,
This element recites insignificant extra solution activity in the form of selecting a particular data source or type of data to be manipulated. It does not integrate the exception.
This element recites a well-understood, routine, and conventional (WURC) computer function (storing and receiving information in memory) and thus fails to recite an inventive concept.
2106.05(g)
2106.05(d)(II)
wherein the input data comprise at least one of: a user request, a document and product-related data,
This element recites insignificant extra solution activity in the form of selecting a particular data source or type of data to be manipulated. It does not integrate the exception.
This element recites a well-understood, routine, and conventional (WURC) computer function (storing and receiving information in memory) and thus fails to recite an inventive concept.
2106.05(g)
2106.05(d)(II)
wherein the metadata comprise at least one of: a user intent and a user sentiment,
This element recites insignificant extra solution activity in the form of selecting a particular data source or type of data to be manipulated. It does not integrate the exception.
This element recites a well-understood, routine, and conventional (WURC) computer function (storing and receiving information in memory) and thus fails to recite an inventive concept.
2106.05(g)
2106.05(d)(II)
and/or wherein the analytical artificial intelligence module, in particular the analytical artificial intelligence metadata generating module comprises:
Anything after “and/or” has no patentable weight because “or” is non-limiting.
11. The system of claim 10,
See claim 10 above.
wherein the first metadata refer to at least one of: the user intent.
This element merely recites the data being evaluated.
2106.04(a)(2)(III)
And wherein the second metadata refer to at least one of content of the input data
This element merely recites the data being evaluated.
2106.04(a)(2)(III)
12. The system of claim 1, further comprising: a search sub-module configured, when executed by at least one of the one or more processors,
This element recite mere instructions to apply the exception and generally links the abstract idea to a computing environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
to search for documents based on the metadata,
This element merely requires observation and evaluation of the metadata and evaluation or judgment as to how to search for the documents based on the metadata.
2106.04(a)(2)(III)
wherein the prompting module is, when executed by at least one of the one or more processors, configured to
This element recite mere instructions to apply the exception and generally links the abstract idea to a computing environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
determine the prompt, based on at least one document found by the search sub-module or a part of the at least one document.
This element merely requires observation and evaluation of the found document and an evaluation/judgment as to how to determine the prompt based on the found document.
2106.04(a)(2)(III)
13. The system of claim 12, wherein the search sub-module is, when executed by at least one of the one or more processors, configured to at least one of:
This element recite mere instructions to apply the exception and generally links the abstract idea to a computing environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
determining, based on the metadata, a weight for any document found by the search sub-module.
This element merely requires observation and evaluation of the found document and a judgment as to how to determine the weight based on the metadata.
2106.04(a)(2)(III)
14. The system of claim 10, wherein the prompting module is, when executed by at least one of the one or more processors, configured to:
See claim 10 and claim 1 above.
determining the prompt based on the first metadata, the second metadata, at least a part of the input data, and at least one document found by the search sub-module or a part thereof; a
This element merely requires evaluation of the first and second metadata and the input data and the found document and an evaluation and/or judgment as to how to determine the prompt.
2106.04(a)(2)(III)
Wherein the system is configured for
See claim 10 and claim 1 above.
searching for the documents based on the first metadata and the second metadata;
This element merely requires evaluation of the first and second metadata and the input data and an evaluation and/or judgment as to how to search for the documents.
2106.04(a)(2)(III)
performing a neural search for documents based on the input data, in particular a question extracted from the input data;
This element merely requires evaluation the input data and an evaluation and/or judgment as to how to search for the documents.
The word “neural” generally links the abstract idea to a neural networks and thus does not integrate or recite an inventive concept.
2106.04(a)(2)(III)
2106.05(f),(h)
and determining, based on the metadata, a weight for any document found by the search sub-module.
This element merely requires observation and evaluation of the metadata and a judgment as to how to determine a weight for the document found by the search sub-module.
2106.04(a)(2)(III)
15. The system of claim 10, wherein the prompting module is, when executed by at least one of the one or more processors, configured to:
This element recite mere instructions to apply the exception and generally links the abstract idea to a computing environment and thus does not integrate or recite an inventive concept.
2106.05(f),(h)
include at least a part of the first metadata, at least a part of the second metadata, at least a part of the input data, and at least a part of at least one document found by the search sub-module into the prompt;
This element merely requires evaluation of the first and second metadata and the input data and the found document and a judgment as to what parts to put in the prompt.
2106.04(a)(2)(III)
Wherein the system is configured for
See claim 10 above and claim 1 above.
searching for the documents based on the first metadata and the second metadata;
This element merely requires evaluation of the first and second metadata and a judgment as to how to search for the documents.
2106.04(a)(2)(III)
performing a neural search for documents based on the input data, in particular a question extracted from the input data;
This element merely requires evaluation the input data and an evaluation and/or judgment as to how to search for the documents.
The word “neural” generally links the abstract idea to neural networks and thus does not integrate or recite an inventive concept.
2106.04(a)(2)(III)
2106.05(f),(h)
and determining, based on the metadata, a weight for any document found by the search sub-module.
This element merely requires observation and evaluation of the metadata and a judgment as to how to determine a weight for the document found by the search sub-module.
2106.04(a)(2)(III)
16. The system of claim 12, wherein the system is configured to at least one of: d
See claim 12 above.
delimit, a number of documents found by the search sub-module to be further used to
This element merely requires evaluation of the found documents and a judgment as to how to delimit them.
2106.04(a)(2)(III)
2106.05(h)
determine at least one of: the prompt and the output data;
This element merely recites the result of the evaluations performed in this claim’s parent claims.
2106.04(a)(2)(III)
and select a subset of the documents based on the weights;
This element merely requires evaluation of the documents and the weights and a judgment as to which documents to select.
2106.04(a)(2)(III)
wherein at least one of: the prompt and the output data is determined based on the subset of the found documents.
This element merely recites the result of the selection (judgment) performed in the element above.
2106.04(a)(2)(III)
17. The system of claim 10, wherein the system is configured to:
See claim 10 above.
feed the first metadata, the second metadata and a subset of documents found by the search sub-module and selected in accordance with corresponding weights to the prompting module;
This element generally links the abstract idea to the field of use of analytical artificial intelligence and generative AI. Thus, it is field of use limitation that does not integrate the exception and does not recite an inventive concept.
2106.05 (h)
and the system being configured for
See claim 10 above.
searching for the documents based on the first metadata and the second metadata;
This element merely requires evaluation of the first and second metadata and a judgment as to how to search for documents based on the first and second metadata.
2106.04(a)(2)(III)
performing a neural search for documents based on the input data, in particular a question extracted from the input data;
This element merely requires evaluation of the input data and an evaluation and/or judgment as to how to search for the documents.
The word “neural” generally links the abstract idea to a neural networks and thus does not integrate or recite an inventive concept.
2106.04(a)(2)(III)
2106.05(f),(h)
and determining, based on the metadata, a weight for any document found by the search sub-module.
This element merely requires observation and evaluation of the metadata and a judgment as to how to determine a weight for the document found by the search sub-module.
2106.04(a)(2)(III)
18. The system of claim 1, wherein the system is configured to at least one of:
See claim 1 above.
output the output data; and display at least a part of the output data, in particular on a display of the system.
These elements generally link the abstract idea to a GUI computer environment and thus does not integrate the exception or recite an inventive concept.
2106.05(h)
The additional elements above “[a]dd nothing … that is not already present when the steps are considered separately’”. MPEP 2106.05 (I)(B)(quoting Alice). As such, when the claim elements above are considered as a whole and individually, claims recite an abstract idea without significantly more.
Claim Rejections - 35 USC § 102
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.
Claim(s) 1-6, 8, and 10-20 is/are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Shen US 20250373576 A1.,
Claim
Shen US 20250373576 A1
Examiner’s comments
1. A system for processing input data, the system comprising: one or more processors; an analytical artificial intelligence module that is, when executed by at least one of the one or more processors, configured to
Fig. 9, 906, 902; Fig. 3 item 305;
determine, metadata for the input data;
[0056] In some embodiments, chat completion 275 generates a response prompt including guidance on style and format. For example, chat completion 275 can determine guidance to include in the response prompt based on metadata of user input
[0065] FIG. 4 illustrates an exemplary user interface 400 in accordance with some embodiments of the present disclosure. As shown in FIG. 4, user interface 400 includes user input 405, response 410, sources 412, and user input interface 415.
[0066] In some embodiments, user interface 400 is implemented on the user interface of a response generation system (e.g., user interface 112 of FIGS. 1 and 2). In response to a user interacting with user input interface 415 and inputting user input 405, the response generation system (e.g., computing system 300 of FIG. 3) determines facets for the user of the user interface and generates response 410. For example, the facets include the kind of product that the user is subscribed to. Accordingly, the response 410 generated and displayed in user interface 400 includes information generated for that specific product. In some embodiments, as shown in FIG. 4, sources 412 is a list of sources included in generated response 410. For example, response 410 includes a list of and/or links to one or more of the relevant content items (e.g., relevant content items 254) used to generate response 410.
[0079] At operation 605, the processing device receives user input from a user system. For example, facet-based response generation component 150 receives user input 242 from user system 110 in response to a user of user system 110 interacting with a chat interface such as user interface 112. In some embodiments, the user input include metadata, such as data about the user of the user system. For example, user input 242 includes data from a profile of a user of user system 110. Further details regarding receiving user input from a user system are discussed with reference to FIG. 2.
Examiner finds metadata includes facets.
and weights of the metadata
Fig. 2 item 254
[0051] In some embodiments, topic classification 285 determines facets for user input 242 using a machine learning model. For example, topic classification 285 applies an LLM to user input 242 to determine facets for user input 242. Content retrieval 270 performs the similarity search on content item embeddings based on the determined facets. For example, content retrieval 270 only performs a similarity search on content item embeddings of content item embeddings 210 with facets that are shared with the determined facets for user input 242. Because performing the similarity search can be a computationally intensive task, computing system 200 saves resources such as computing power and time by only performing the similarity search on a subset of content item embeddings 210. Accordingly, the total throughput of computing system 200 is improved as a result of using this facet-based exclusion.
.
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Subset of found documents returned are based on weight in that the embedding of the facets provide the weight. Examiner finds facets are weighted positively because they are similar. Similarity threshold is a type of weight.
a prompting module that is, when executed by at least one of the one or more processors, configured to determine, based at least on a part of the metadata, and the weights of the metadata a prompt;
[0056] In some embodiments, chat completion 275 generates a response prompt including guidance on style and format. For example, chat completion 275 can determine guidance to include in the response prompt based on metadata of user input
[0070] At operation 515, the processing device matches user input embedding to a set of content items. For example, facet labeling component 160 determines relevant content items 254 by performing a similarity search using user input embedding 252 and content item embeddings 210. In some embodiments, the processing device uses facet-based exclusion to determine a set of content item embeddings in addition to performing the similarity search. For example, facet labeling component 160 determines facets for user input 242 and generates relevant content items 254 by performing a similarity search on content item embeddings 210 with facets that match user input 242. Further details regarding matching user input to a set of content items are discussed with reference to FIGS. 2 and 5.
[0071] At operation 520, the processing device generates a response to the user input using the set of content items. For example, facet labeling component 160 generates a response prompt using user input embedding 252 and relevant content items 254. Facet labeling component 160 applies the response prompt to a generative machine learning model to generate response 258 using relevant content items 254. Further details regarding generating a response to the user input using the set of content items are discussed with reference to FIGS. 2 and 6.
[0081] At operation 615, the processing device creates a standardized prompt using the user input. For example, facet-based response generation component 150 creates a prompt including the user input (e.g., text) from the user interacting with user interface 112. In some embodiments, the processing device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to. Further details regarding creating a standardized prompt using the user input are discussed with reference to FIG. 2.
Examiner finds the facets, that are part of the metadata, cause the metadata to be weighted a particular way. For example, “. . . the device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to.”
and a generative artificial intelligence module that is, when executed by at least one of the one or more processors, configured to determine, based on the prompt, output data.
[0057] Chat completion 275 applies a machine learning model (e.g., generative machine learning model component 305 of FIG. 3) to the generated response prompt, causing the machine learning model to generate response candidate 256. For example, chat completion 275 sends the response prompt to a generative machine learning model which generates response candidate 256 based on the response prompt (e.g., a response to user input 242 based on accessing relevant content items 254). In some embodiments, the generative machine learning model is provided access to relevant content items 254 but only uses a subset of relevant content items 254 in generating response candidate 256. Chat completion 275 sends response candidate 256 to response validation 280.
[0071] At operation 520, the processing device generates a response to the user input using the set of content items. For example, facet labeling component 160 generates a response prompt using user input embedding 252 and relevant content items 254. Facet labeling component 160 applies the response prompt to a generative machine learning model to generate response 258 using relevant content items 254. Further details regarding generating a response to the user input using the set of content items are discussed with reference to FIGS. 2 and 6.
Response and/or response candidate is the output data
19. A computer-implemented method for improving human-machine interaction, the method comprising: feeding input data to an analytical artificial intelligence module for determining metadata for the input data;
Fig. 2 item 242, Fig. 3 item 305; Fig. 3 item 150; Fig. 2 items 240, 245
[0044] User input standardization 240 receives user input 242 from user system 110 and processes user input 242 into standardized user input 244. For example, user input standardization 240 processes user input 242 into a standardized search query format (e.g., how to). In some embodiments, user input standardization 240 generates standardized user input 244 including a prompt for a machine learning model. For example, user input standardization 240 generates a prompt (e.g., standardized user input 244) for user input 242 including instructions to generate a standardized search query for user input 242. In some embodiments, user input standardization 240 processes user input 242 into a standardized format using metadata of user input 242. For example, user input standardization 240 generates a prompt for user input 242 with instructions that are based on the metadata associated with user input 242. In one embodiment, user input standardization 240 generates standardized user input 244 including a prompt with instructions based on a product the user of user system 110 is subscribed to. . . .
[0045] Intent classification 245 receives standardized user input 244 and generates user intent 246. For example, intent classification 245 classifies the intent of standardized user input 244. The intent can include, for example, whether the user input includes a desire to speak with an agent, whether the user input includes a greeting, whether the user input includes a prompt injection, whether the user input is requesting help, etc. In some embodiments, intent classification 245 generates user intent 246 using user input 242. In some embodiments, intent classification 245 generates user intent 246 using standardized user input 244. In some embodiments, intent classification 245 generates user intent 246 using metadata of user input 242 and/or standardized user input 244. For example, the metadata of user input 242 and/or standardized user input 244 includes historical data indicating that the user has recently performed a search for how to address an account problem. In such an example, intent classification 245 can determine a user intent 246 for help based on this historical data. Intent classification 245 sends user intent 246 to search query refining 255.
[0079] At operation 605, the processing device receives user input from a user system. For example, facet-based response generation component 150 receives user input 242 from user system 110 in response to a user of user system 110 interacting with a chat interface such as user interface 112. In some embodiments, the user input include metadata, such as data about the user of the user system. For example, user input 242 includes data from a profile of a user of user system 110. Further details regarding receiving user input from a user system are discussed with reference to FIG. 2.
User input 242 is input data that is feed into item 150 which is part of AI intelligence modules (e.g. Fig. 3 item 300).
User input is used to determine metadata of user input.
Examiner finds Fig. 3 item 305 performs analytics and thus teaches analytical artificial intelligence module (logic) that performs said analytics.
And the weights of the metadata
Fig. 2 item 254
[0051] In some embodiments, topic classification 285 determines facets for user input 242 using a machine learning model. For example, topic classification 285 applies an LLM to user input 242 to determine facets for user input 242. Content retrieval 270 performs the similarity search on content item embeddings based on the determined facets. For example, content retrieval 270 only performs a similarity search on content item embeddings of content item embeddings 210 with facets that are shared with the determined facets for user input 242. Because performing the similarity search can be a computationally intensive task, computing system 200 saves resources such as computing power and time by only performing the similarity search on a subset of content item embeddings 210. Accordingly, the total throughput of computing system 200 is improved as a result of using this facet-based exclusion.
.
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Subset of found documents returned are based on weight in that the embedding of the facets provide the weight. Examiner finds facets are weighted positively because they are similar. Similarity threshold is a type of weight.
feeding at least a part of the metadata to a prompting module for determining a prompt for the input data;
Fig. 2 items 245, 240, 270, 275
[0044] User input standardization 240 receives user input 242 from user system 110 and processes user input 242 into standardized user input 244. For example, user input standardization 240 processes user input 242 into a standardized search query format (e.g., how to). In some embodiments, user input standardization 240 generates standardized user input 244 including a prompt for a machine learning model. For example, user input standardization 240 generates a prompt (e.g., standardized user input 244) for user input 242 including instructions to generate a standardized search query for user input 242. In some embodiments, user input standardization 240 processes user input 242 into a standardized format using metadata of user input 242. For example, user input standardization 240 generates a prompt for user input 242 with instructions that are based on the metadata associated with user input 242. In one embodiment, user input standardization 240 generates standardized user input 244 including a prompt with instructions based on a product the user of user system 110 is subscribed to. . . .
[0046] Search query refining 255 receives user intent 246 and generates refined search query 250 using user intent 246 and standardized user input 244. For example, search query refining 255 generates refined search query 250 including a prompt including the prompt generated by user input standardization 240 (e.g., standardized user input 244) and the user intent 246 determined by intent classification 245.
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Items 244, 246, 250, 252, 254 fed into 275 and used to determine a response prompt.
the prompt including metadata which are at least one of ranked in accordance with the weights of the metadata and selected based on the weights of the metadata; and
[0056] In some embodiments, chat completion 275 generates a response prompt including guidance on style and format. For example, chat completion 275 can determine guidance to include in the response prompt based on metadata of user input
[0070] At operation 515, the processing device matches user input embedding to a set of content items. For example, facet labeling component 160 determines relevant content items 254 by performing a similarity search using user input embedding 252 and content item embeddings 210. In some embodiments, the processing device uses facet-based exclusion to determine a set of content item embeddings in addition to performing the similarity search. For example, facet labeling component 160 determines facets for user input 242 and generates relevant content items 254 by performing a similarity search on content item embeddings 210 with facets that match user input 242. Further details regarding matching user input to a set of content items are discussed with reference to FIGS. 2 and 5.
[0071] At operation 520, the processing device generates a response to the user input using the set of content items. For example, facet labeling component 160 generates a response prompt using user input embedding 252 and relevant content items 254. Facet labeling component 160 applies the response prompt to a generative machine learning model to generate response 258 using relevant content items 254. Further details regarding generating a response to the user input using the set of content items are discussed with reference to FIGS. 2 and 6.
[0081] At operation 615, the processing device creates a standardized prompt using the user input. For example, facet-based response generation component 150 creates a prompt including the user input (e.g., text) from the user interacting with user interface 112. In some embodiments, the processing device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to. Further details regarding creating a standardized prompt using the user input are discussed with reference to FIG. 2.
Examiner finds the facets, that are part of the metadata, cause the metadata to be weighted (ranked) a particular way. This weighted metadata is used to generate prompts. For example, “. . . the device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to.”
and feeding the prompt to a generative artificial intelligence module for determining output data for the input data.
Fig. 2 item 275, 256, 280, 258
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
0057] Chat completion 275 applies a machine learning model (e.g., generative machine learning model component 305 of FIG. 3) to the generated response prompt, causing the machine learning model to generate response candidate 256. For example, chat completion 275 sends the response prompt to a generative machine learning model which generates response candidate 256 based on the response prompt (e.g., a response to user input 242 based on accessing relevant content items 254). In some embodiments, the generative machine learning model is provided access to relevant content items 254 but only uses a subset of relevant content items 254 in generating response candidate 256. Chat completion 275 sends response candidate 256 to response validation 280.
0058] Response validation 280 receives response candidate 256 and determines whether to send response candidate 256 as response 258. For example, response validation 280 checks for hallucinations and/or inappropriate content (e.g., responses including profanity, legal content, and/or prejudicial content) and sends response 258 in response to successfully validating response candidate 256. In some embodiments, response validation 280 determines whether to send response candidate 256 as response 258 using a machine learning model. For example, response validation 280 generates a prompt for a generative machine learning model to determine whether response candidate 256 includes hallucinations and/or inappropriate content. Facet-based response generation component 150 sends response 258 to user system 110. For example, facet-based response generation component 150 sends response 258 to user system 110 via a chat interface of user interface 112, causing user interface 112 to display response 258. In some embodiments, response validation 280 stores response 258 in chat history 235 for future access.
Prompt (including input data) fed into generative AI for output data (response).
20. A computer program product or a non-volatile computer-readable medium comprising instructions which, when executed by one or more processors of a system, cause the system to carry out the following processes: feeding input data to an analytical artificial intelligence module for determining metadata for the input data;
Fig. 2 item 242, Fig. 3 item 305; Fig. 3 item 150; Fig. 2 items 240, 245
[0044] User input standardization 240 receives user input 242 from user system 110 and processes user input 242 into standardized user input 244. For example, user input standardization 240 processes user input 242 into a standardized search query format (e.g., how to). In some embodiments, user input standardization 240 generates standardized user input 244 including a prompt for a machine learning model. For example, user input standardization 240 generates a prompt (e.g., standardized user input 244) for user input 242 including instructions to generate a standardized search query for user input 242. In some embodiments, user input standardization 240 processes user input 242 into a standardized format using metadata of user input 242. For example, user input standardization 240 generates a prompt for user input 242 with instructions that are based on the metadata associated with user input 242. In one embodiment, user input standardization 240 generates standardized user input 244 including a prompt with instructions based on a product the user of user system 110 is subscribed to. . . .
[0045] Intent classification 245 receives standardized user input 244 and generates user intent 246. For example, intent classification 245 classifies the intent of standardized user input 244. The intent can include, for example, whether the user input includes a desire to speak with an agent, whether the user input includes a greeting, whether the user input includes a prompt injection, whether the user input is requesting help, etc. In some embodiments, intent classification 245 generates user intent 246 using user input 242. In some embodiments, intent classification 245 generates user intent 246 using standardized user input 244. In some embodiments, intent classification 245 generates user intent 246 using metadata of user input 242 and/or standardized user input 244. For example, the metadata of user input 242 and/or standardized user input 244 includes historical data indicating that the user has recently performed a search for how to address an account problem. In such an example, intent classification 245 can determine a user intent 246 for help based on this historical data. Intent classification 245 sends user intent 246 to search query refining 255.
[0079] At operation 605, the processing device receives user input from a user system. For example, facet-based response generation component 150 receives user input 242 from user system 110 in response to a user of user system 110 interacting with a chat interface such as user interface 112. In some embodiments, the user input include metadata, such as data about the user of the user system. For example, user input 242 includes data from a profile of a user of user system 110. Further details regarding receiving user input from a user system are discussed with reference to FIG. 2.
User input 242 is input data that is feed into item 150 which is part of AI intelligence modules (e.g. Fig. 3 item 300).
User input is used to determine metadata of user input.
Examiner finds Fig. 3 item 305 performs analytics and thus teaches analytical artificial intelligence module (logic) that performs said analytics.
feeding at least a part of the metadata to a prompting module for determining a prompt for the input data;
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Metadata related to items 242, 244, 246, 250, 252, 254 fed into 275 and used to determine a response prompt for user input 242.
And weights of the metadata
Fig. 2 item 254; para. 50;
[0051] In some embodiments, topic classification 285 determines facets for user input 242 using a machine learning model. For example, topic classification 285 applies an LLM to user input 242 to determine facets for user input 242. Content retrieval 270 performs the similarity search on content item embeddings based on the determined facets. For example, content retrieval 270 only performs a similarity search on content item embeddings of content item embeddings 210 with facets that are shared with the determined facets for user input 242. Because performing the similarity search can be a computationally intensive task, computing system 200 saves resources such as computing power and time by only performing the similarity search on a subset of content item embeddings 210. Accordingly, the total throughput of computing system 200 is improved as a result of using this facet-based exclusion.
.
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Subset of found documents returned are based on weight in that the embedding of the facets provide the weight. Examiner finds facets are weighted positively because they are similar. Similarity threshold is a type of weight.
and feeding the prompt to a generative artificial intelligence module for determining output data for the input data
[0057] Chat completion 275 applies a machine learning model (e.g., generative machine learning model component 305 of FIG. 3) to the generated response prompt, causing the machine learning model to generate response candidate 256. For example, chat completion 275 sends the response prompt to a generative machine learning model which generates response candidate 256 based on the response prompt (e.g., a response to user input 242 based on accessing relevant content items 254). In some embodiments, the generative machine learning model is provided access to relevant content items 254 but only uses a subset of relevant content items 254 in generating response candidate 256. Chat completion 275 sends response candidate 256 to response validation 280.
[0058] Response validation 280 receives response candidate 256 and determines whether to send response candidate 256 as response 258. For example, response validation 280 checks for hallucinations and/or inappropriate content (e.g., responses including profanity, legal content, and/or prejudicial content) and sends response 258 in response to successfully validating response candidate 256. In some embodiments, response validation 280 determines whether to send response candidate 256 as response 258 using a machine learning model. For example, response validation 280 generates a prompt for a generative machine learning model to determine whether response candidate 256 includes hallucinations and/or inappropriate content. Facet-based response generation component 150 sends response 258 to user system 110. For example, facet-based response generation component 150 sends response 258 to user system 110 via a chat interface of user interface 112, causing user interface 112 to display response 258. In some embodiments, response validation 280 stores response 258 in chat history 235 for future access.
275 feeds prompt into generative AI module (generative machine learning model component 305) for determining output (responses) for user input (input data).
the prompt including metadata which are at least one of ranked in accordance with the weights of the metadata and selected based on the weights of the metadata; and
para. 50;
[0056] In some embodiments, chat completion 275 generates a response prompt including guidance on style and format. For example, chat completion 275 can determine guidance to include in the response prompt based on metadata of user input
[0070] At operation 515, the processing device matches user input embedding to a set of content items. For example, facet labeling component 160 determines relevant content items 254 by performing a similarity search using user input embedding 252 and content item embeddings 210. In some embodiments, the processing device uses facet-based exclusion to determine a set of content item embeddings in addition to performing the similarity search. For example, facet labeling component 160 determines facets for user input 242 and generates relevant content items 254 by performing a similarity search on content item embeddings 210 with facets that match user input 242. Further details regarding matching user input to a set of content items are discussed with reference to FIGS. 2 and 5.
[0071] At operation 520, the processing device generates a response to the user input using the set of content items. For example, facet labeling component 160 generates a response prompt using user input embedding 252 and relevant content items 254. Facet labeling component 160 applies the response prompt to a generative machine learning model to generate response 258 using relevant content items 254. Further details regarding generating a response to the user input using the set of content items are discussed with reference to FIGS. 2 and 6.
[0081] At operation 615, the processing device creates a standardized prompt using the user input. For example, facet-based response generation component 150 creates a prompt including the user input (e.g., text) from the user interacting with user interface 112. In some embodiments, the processing device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to. Further details regarding creating a standardized prompt using the user input are discussed with reference to FIG. 2.
Examiner finds the facets, that are part of the metadata, cause the metadata to be weighted a particular way. The weighted (ranked metadata) is used to generate prompts. For example, “. . . the device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to.”
and feeding the prompt to a generative artificial intelligence module for determining output data for the input data.
Fig. 2 item 275, 256, 280, 258
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
0057] Chat completion 275 applies a machine learning model (e.g., generative machine learning model component 305 of FIG. 3) to the generated response prompt, causing the machine learning model to generate response candidate 256. For example, chat completion 275 sends the response prompt to a generative machine learning model which generates response candidate 256 based on the response prompt (e.g., a response to user input 242 based on accessing relevant content items 254). In some embodiments, the generative machine learning model is provided access to relevant content items 254 but only uses a subset of relevant content items 254 in generating response candidate 256. Chat completion 275 sends response candidate 256 to response validation 280.
0058] Response validation 280 receives response candidate 256 and determines whether to send response candidate 256 as response 258. For example, response validation 280 checks for hallucinations and/or inappropriate content (e.g., responses including profanity, legal content, and/or prejudicial content) and sends response 258 in response to successfully validating response candidate 256. In some embodiments, response validation 280 determines whether to send response candidate 256 as response 258 using a machine learning model. For example, response validation 280 generates a prompt for a generative machine learning model to determine whether response candidate 256 includes hallucinations and/or inappropriate content. Facet-based response generation component 150 sends response 258 to user system 110. For example, facet-based response generation component 150 sends response 258 to user system 110 via a chat interface of user interface 112, causing user interface 112 to display response 258. In some embodiments, response validation 280 stores response 258 in chat history 235 for future access.
Prompt (including input data) fed into generative AI for output data (response).
2. The system of claim 1, wherein the analytical artificial intelligence module comprises:
at least one preprocessing module that is, when executed by at least one of the one or more processors, configured to at least one of:
linguistically process the input data;
[0016]. . .The response generation can use data associated with user input (e.g., facets and/or intent of the user input) to filter the content item embeddings available when generating the response to the user input. . . .
[0045] Intent classification 245 receives standardized user input 244 and generates user intent 246. For example, intent classification 245 classifies the intent of standardized user input 244. The intent can include, for example, whether the user input includes a desire to speak with an agent, whether the user input includes a greeting, whether the user input includes a prompt injection, whether the user input is requesting help, etc.
Examiner finds determining intent of user is to linguistically process the input of the user.
determine, for the input data, a metadata request;
[0043] In some embodiments, facet-based response generation component 150 retrieves the data in response to receiving user input 242. For example, user input 242 includes an identifier for the user of user system 110 and facet-based response generation component 150 retrieves data for that user from a data store (e.g., data store 140 of FIG. 1) in response to receiving user input 242.
Metadata request is determined based on user input (e.g. an identifier).
an analytical artificial intelligence metadata generating module that is, when executed by at least one of the one or more processors, configured to at least one of:
determining the respective metadata in accordance with the respective metadata request;
[0043] In some embodiments, facet-based response generation component 150 retrieves the data in response to receiving user input 242. For example, user input 242 includes an identifier for the user of user system 110 and facet-based response generation component 150 retrieves data for that user from a data store (e.g., data store 140 of FIG. 1) in response to receiving user input 242.
Metadata retrieved for user based on identifier.
3. The system of claim 1, wherein the metadata comprise at least one of: . . . category of the content or a topic of the content
0050] In some embodiments, topic classification 285 determines relevant content items 254 of content item embeddings 210 using facet-based exclusion. For example, topic classification 285 determines facets for user input 242. In some embodiments, topic classification 285 determines the facets based on metadata of user input 242. For example, topic classification 285 determines the facets based on products to which the user of user system 110 is subscribed and/or an access level for the user of user system 110. In such embodiments, topic classification 285 can determine a set of content items embeddings to use from content item embeddings 210 stored in vector store 230 based on these determined facets. For example, topic classification 285 can exclude all content items that do not have facets that match the determined facets for user input 242.
Topic classification teaches topic/category of the content. Product metadata teaches content metadata.
4. The system of claim 1, wherein the system is a hybrid artificial intelligence system, wherein the system is a heterogeneous artificial intelligence system,
Fig. 1 item 100; Para. 29, para. 33, para. 39, para. 44, and para. 51, and para. 57, and para. 61.
System 100 includes hybrid/heterogenous AI system because it uses various ML models to perform different tasks such, as generate embeddings, generate responses, and determine facets, among other things.
wherein the input data comprise input textual data, wherein the output data comprise output textual data,
Fig. 4 items 405 and 410 and 412
Input data is item 405; output data is items 410 and 412; these are textual data
wherein the generative artificial intelligence module is based on machine learning, implemented in a container, configured to generate the output text data, and/or comprises a Large Language Model,
Abstract, Fig. 4; para. 62
Fig. 4 is an LLM/generative ML model that generates output text.
and wherein the analytical artificial intelligence module is based on machine learning, implemented in a container, and/or comprises at least one of: an intent analysis capability
Abstract, Fig. 4; para. 62; para. 16; para. 31
Fig. 4 is an based on machine learning; intent classification is intent analysis
5. The system of claim 1, wherein the analytical artificial intelligence module is, when executed by at least one of the one or more processors, configured to determine, based on the metadata for the input data, a request for at least one of: a user feedback and further input data.
Fig. 2 items 235, 258, and 242
[0047] In some embodiments, search query refining 255 receives chat history 235. For example, chat history 235 includes data about previous interactions between the user or the user system 110 and facet-based response generation component 150. Chat history 235 can include, for example, previous requests (e.g., user inputs) sent by the user or the user system 110 and responses to those previous requests sent by facet-based response generation component 150. In such embodiments, search query refining 255 generates refined search query 250 based on the context provided by chat history 235. For example, refined search query 250 can include a prompt with a statement indicating previous unsuccessful responses and/or previous information provided by user system 110. Search query refining 255 sends refined search query 250 to user input embedding component 265.
User input data includes metadata that requests user feedback and/or further input data in the form of the user’s previous chat history.
6. The system of claim 5, wherein the analytical artificial intelligence module is, when executed by at least one of the one or more processors, configured to update the metadata in accordance with a received user feedback and/or based on received further input data.
Fig. 2 items 235, 258, and 242
[0047] In some embodiments, search query refining 255 receives chat history 235. For example, chat history 235 includes data about previous interactions between the user or the user system 110 and facet-based response generation component 150. Chat history 235 can include, for example, previous requests (e.g., user inputs) sent by the user or the user system 110 and responses to those previous requests sent by facet-based response generation component 150. In such embodiments, search query refining 255 generates refined search query 250 based on the context provided by chat history 235. For example, refined search query 250 can include a prompt with a statement indicating previous unsuccessful responses and/or previous information provided by user system 110. Search query refining 255 sends refined search query 250 to user input embedding component 265.
Chat history (metadata is updated) each time user chats with system. Examiner finds this is feedback and/or further input data.
8. The system of claim 1, wherein the prompting module is, when executed by at least one of the one or more processors, configured to at least one of: selecting based on the weights of the metadata, metadata to be included into the prompt;
[0080] At operation 610, the processing device determines a set of facets using the user input. For example, facet-based response generation component 150 determines a set of facets for user input 242 based on the content of user input 242. In some embodiments, facet-based response generation component 150 determines the set of facets based on metadata of user input 242. For example, facet-based response generation component 150 determines the set of facets based on data taken from the profile of the user of user system 110 such as products that the user is subscribed to. Further details regarding determining a set of facets using the user input are discussed with reference to FIG. 2.
[0081] At operation 615, the processing device creates a standardized prompt using the user input. For example, facet-based response generation component 150 creates a prompt including the user input (e.g., text) from the user interacting with user interface 112. In some embodiments, the processing device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to.
Metadata such as products that user is subscribed it is included in the prompt. The generation of the prompt is based on weights in that the facets weigh the metadata in a particular way (e.g. products user is subscribed to).
10. (Currently Amended) The system of claim 1, wherein the input data are received via an interface from a user,
Para. 44 Fig. 2 items 242 and 244;
wherein the input data comprise at least one of: a user request, a document and product-related data,
Para. 42; Fig. 2 items 242
User input 242 includes at least product related data, i.e. “data about the products that the user is subscribed to”
wherein the metadata comprise at least one of: a user intent and a user sentiment, and/or
Para. 52
Metadata includes user intent. Examiner finds prior art is not required to teach anything after “and/or” because “or” is non-limiting.
11. The system of claim 10, wherein the first metadata refer to at least one of: the user intent.
0016]. . .The response generation can use data associated with user input (e.g., facets and/or intent of the user input) to filter the content item embeddings available when generating the response to the user input. . . .
[0045] Intent classification 245 receives standardized user input 244 and generates user intent 246. For example, intent classification 245 classifies the intent of standardized user input 244. The intent can include, for example, whether the user input includes a desire to speak with an agent, whether the user input includes a greeting, whether the user input includes a prompt injection, whether the user input is requesting help, etc.
User intent includes desire to speak to agent, greeting, injection, help, etc.
12. The system of claim 1, further comprising: a search sub-module configured, when executed by at least one of the one or more processors, to search for documents based on the metadata,
Fig. 2 item 270, 252; para. 32, para. 49
A search for content items (documents) happens based on user input which includes metadata.
wherein the prompting module is, when executed by at least one of the one or more processors, configured to determine the prompt, based on at least one document found by the search sub-module or a part of the at least one document.
Fig. 2 item 275;
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Content items (documents) used to generate (determine) the prompt.
13. The system of claim 12, wherein the search sub-module is, when executed by at least one of the one or more processors, configured to at least one of: determining, based on the metadata, a weight for any document found by the search sub-module.
Para. 31,
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Vector encoding puts a weight on documents in that they are filtered out based on their similarity to the user’s input and metadata based on the vector encoding.
Examiner finds non-relevant documents (content items) are effectively weigh zero and relevant documents have a non-zero, positive weight.
14. The system of claim 10, wherein the prompting module is, when executed by at least one of the one or more processors, configured to: determining the prompt based on the first metadata, the second metadata, at least a part of the input data, and at least one document found by the search sub-module or a part thereof; a
Fig. 2 item 240, 245, 270 and 275; paras. 44, 45, 49
Metadata corresponding to input standardization and intent classification and content items (documents) retrieved by content retrieval are all used to generate (determine) the prompt.
and wherein the system is configured for searching for the documents based on the first metadata and the second metadata;
Para. 49, Fig. 2 items 240, 245, 255, 265, 270, and 275;
Content retrieval searches the documents (content items) based on metadata corresponding to input standardization 240 and intent classification 245, for example.
performing a neural search for documents based on the input data, in particular a question extracted from the input data;
Fig. 2 item 242, 240, 244, 270, Fig. 4 item 405; para. 49 and para. 61 (neural network technology)
Fig. 4 item 405 is a question extracted from user input; neural search happens because the search is based on neural network technology.
and determining, based on the metadata, a weight for any document found by the search sub-module.
Paras. 31, 39, 49, and 51
Weight of document is determined based on the vector embedding of its facet and the facets’ similarities to user input and metadata
15. The system of claim 10, wherein the prompting module is, when executed by at least one of the one or more processors, configured to: include at least a part of the first metadata, at least a part of the second metadata, at least a part of the input data, and at least a part of at least one document found by the search sub-module into the prompt;
[0046] Search query refining 255 receives user intent 246 and generates refined search query 250 using user intent 246 and standardized user input 244. For example, search query refining 255 generates refined search query 250 including a prompt including the prompt generated by user input standardization 240 (e.g., standardized user input 244) and the user intent 246 determined by intent classification 245.
Para. 48
For example, user input embedding component 265 creates a prompt instructing a machine learning model to generate an embedding for refined search query 250. User input embedding component 265 applies the generated prompt to the generative machine learning model causing the generative machine learning model to generate user input embedding 252. User input embedding component 265 sends user input embedding 252 to content retrieval 270.
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
Prompt is created based on at least part of the metadata associated with user standardization and user intent and relevant content items (documents) found by content retrieval 270.
and the system further comprising: searching for the documents based on the first metadata and the second metadata;
[0046] Search query refining 255 receives user intent 246 and generates refined search query 250 using user intent 246 and standardized user input 244. For example, search query refining 255 generates refined search query 250 including a prompt including the prompt generated by user input standardization 240 (e.g., standardized user input 244) and the user intent 246 determined by intent classification 245.
Content searched based on metadata corresponding to 240 and 246.
performing a neural search for documents based on the input data, in particular a question extracted from the input data;
Fig. 2 item 242; Fig. 4 item 405;
[0046] Search query refining 255 receives user intent 246 and generates refined search query 250 using user intent 246 and standardized user input 244. For example, search query refining 255 generates refined search query 250 including a prompt including the prompt generated by user input standardization 240 (e.g., standardized user input 244) and the user intent 246 determined by intent classification 245.
[0048] User input embedding component 265 receives refined search query 250 and generates user input embedding 252 using refined search query 250.
[0049] Content retrieval 270 retrieves relevant content items 254 using user input embedding 252. For example, content retrieval 270 performs a similarity search between the content item embeddings 210 of vector store 230 and user input embedding 252 and retrieves relevant content items 254 based on the content item embeddings with a high degree of similarity to user input embedding 252. In some embodiments, content retrieval 270 determines relevant content items 254 as the content items with embeddings that have the highest similarity to user input embedding 252. For example, content retrieval 270 determines relevant content items 254 and the content items with embeddings that are the top ten most similar (e.g., shortest distance in the vector space of user input embedding 252 and content item embeddings 210) to user input embedding 252. In some embodiments content retrieval 270 determines relevant content items 254 based on a similarity threshold. For example, content retrieval 270 determines relevant content items 254 using embeddings with similarity search results that satisfy a similarity threshold (e.g., a certain distance in the vector space and/or a certain percent similarity). In some embodiments, content retrieval 270 performs a similarity search using a cosine similarity search.
Question is extracted at least via 240 and 245.
and determining, based on the metadata, a weight for any document found by the search sub-module.
Paras. 31, 39, 49, and 51
Weight of document is determined based on the vector embedding of its facet and the facets’ similarities to user input and metadata
16. The system of claim 12, wherein the system is configured to at least one of: delimit, a number of documents found by the search sub-module to be further used to determine at least one of:
[0051] In some embodiments, topic classification 285 determines facets for user input 242 using a machine learning model. For example, topic classification 285 applies an LLM to user input 242 to determine facets for user input 242. Content retrieval 270 performs the similarity search on content item embeddings based on the determined facets. For example, content retrieval 270 only performs a similarity search on content item embeddings of content item embeddings 210 with facets that are shared with the determined facets for user input 242. Because performing the similarity search can be a computationally intensive task, computing system 200 saves resources such as computing power and time by only performing the similarity search on a subset of content item embeddings 210. Accordingly, the total throughput of computing system 200 is improved as a result of using this facet-based exclusion.
[0057] Chat completion 275 applies a machine learning model (e.g., generative machine learning model component 305 of FIG. 3) to the generated response prompt, causing the machine learning model to generate response candidate 256. For example, chat completion 275 sends the response prompt to a generative machine learning model which generates response candidate 256 based on the response prompt (e.g., a response to user input 242 based on accessing relevant content items 254). In some embodiments, the generative machine learning model is provided access to relevant content items 254 but only uses a subset of relevant content items 254 in generating response candidate 256. Chat completion 275 sends response candidate 256 to response validation 280.
Examiner finds a subset of documents teaches “delimiting” a number of documents.
the prompt and the output data;
[0054] As mentioned above, performing a similarity search can be a resource intensive task. Since content item embeddings are stored in chunks in vector store 230 (e.g., smaller portions of the whole content item), content retrieval 270 can more quickly determine similarity between user input embedding 252 and a chunk of a content item than if the content item embedding were stored in its entirety. In response to determining that a chunk of a content item is relevant (e.g., satisfies the similarity threshold), content retrieval 270 can then retrieve the other chunks of the content item using the metadata without the need to perform a resource intensive similarity search on the content item as a whole. Accordingly, by using chunked content item embeddings, computing system 200 saves computing power and time and increases the throughput of the system as a whole. Content retrieval 270 sends relevant content items 254 to chat completion 275.
[0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
and select a subset of the documents based on the weights;
[0051] In some embodiments, topic classification 285 determines facets for user input 242 using a machine learning model. For example, topic classification 285 applies an LLM to user input 242 to determine facets for user input 242. Content retrieval 270 performs the similarity search on content item embeddings based on the determined facets. For example, content retrieval 270 only performs a similarity search on content item embeddings of content item embeddings 210 with facets that are shared with the determined facets for user input 242. Because performing the similarity search can be a computationally intensive task, computing system 200 saves resources such as computing power and time by only performing the similarity search on a subset of content item embeddings 210. Accordingly, the total throughput of computing system 200 is improved as a result of using this facet-based exclusion
Subset is selected based on weight in that the embedding of the facets provide the weight. Examiner finds similar facets have a positive weight because they satisfy a threshold.
wherein at least one of: the prompt and the output data is determined based on the subset of the found documents.
0055] Chat completion 275 generates a response prompt using user input embedding 252 and relevant content items 254. For example, chat completion 275 generates a prompt instructing a machine learning model to generate a response to user input 242 represented by user input embedding 252 using resources from relevant content items 254. By providing only relevant content items 254 (e.g., based on facet-based exclusion and similarity search), computing system 200 can prevent hallucination in the response generated by the machine learning model. For example, because content items can include semantically similar material for different products, a system that does not use facet-based exclusion could generate a response to mixes instructions for multiple products, creating a response to does not address the problems for users of either product (or only addresses the problems for a single product). By using facet-based exclusion (e.g., based on metadata associated with the user of user system 110), computing system 200 can restrict the content items consulted when generating response candidate 256 forcing the machine learning model to only rely upon relevant data and thereby preventing hallucination.
The subset of documents are used in that only the relevant ones (similar ones) are used. That is, the ones with similarity encoded facets.
17. The system of claim 10, wherein the system is configured to: feed the first metadata, the second metadata a subset of documents found by the search sub-module and selected in accordance with corresponding weights to the prompting module; and the system being configured for
Fig. 2 items 240, 245, 270, and 725; para. 49
Content retrieval includes finding documents. The documents are weighted in that their facets determine whether they are relevant. That is, the similarity threshold is a weight.
searching for the documents based on the first metadata and the second metadata;
Fig. 2 item 240, 245, 255, 265, 270, 254; para. 48=49
Content retrieval returns relevant content items (documents) based on at least on metadata corresponding to 240 and 245 because they are encoded in the user input embedding 265.
performing a neural search for documents based on the input data, in particular a question extracted from the input data;
Fig. 2 item 242, 110, 112, Fig. 4 item 405
and determining, based on the metadata, a weight for any document found by the search sub-module.
Fig. 2 item 254
[0051] In some embodiments, topic classification 285 determines facets for user input 242 using a machine learning model. For example, topic classification 285 applies an LLM to user input 242 to determine facets for user input 242. Content retrieval 270 performs the similarity search on content item embeddings based on the determined facets. For example, content retrieval 270 only performs a similarity search on content item embeddings of content item embeddings 210 with facets that are shared with the determined facets for user input 242. Because performing the similarity search can be a computationally intensive task, computing system 200 saves resources such as computing power and time by only performing the similarity search on a subset of content item embeddings 210. Accordingly, the total throughput of computing system 200 is improved as a result of using this facet-based exclusion.
Subset of found documents returned based on weight in that the embedding of the facets provide the weight. Examiner finds facets are weighted positively because they are similar. Similarity threshold is a type of weight.
18. The system of claim 1, wherein the system is configured to at least one of: output the output data; and display at least a part of the output data, in particular on a display of the system.
Fig. 4 items 410, 412, 400; Fig. 2 item 110; 112
410 and 412 are output data displayed on 400.
Response to Arguments
Applicant argues
Applicants respectfully disagree and submit that claim 1 is not merely directed to the abstract idea of concepts performed in the human mind (i.e., mental steps to observe, evaluate, and make judgements about electronic data and mathematical calculations). Instead, independent claim 1 introduces to the field of processing input data the novel system of using a machine-based analytical artificial intelligence module to determine metadata for the input data and weights of the metadata, determining a prompt based on the metadata and the weights of the metadata acquired via the machine-based artificial intelligence model, and using a machine-based generative artificial intelligence module to provide output data based on the prompt. Such system and method goes well beyond human mental processes or human judgement and, in fact, eliminates human judgment to provide an objective, data-driven approach to provide improved output from input data.\
“Claims do not recite a mental process when they do not contain limitations that can practically be performed in the human mind, for instance when the human mind is not equipped to perform the claim limitations.” MPEP 2106.04(a)(1)(III)(A).
There is nothing in the following language that cannot be practically performed in the human mind:
determine metadata for the input data and weights of the metadata, determining a prompt based on the metadata and the weights of the metadata and
determine, based at least on a part of the metadata and the weights of the metadata, a prompt
Determining metadata weights and determining a prompt using metadata weights merely requires evaluation and/or judgment as to how much weight to give the metadata and an evaluation/judgment as to how to craft the prompt according to the weights. The recited language above does not reflect Applicant’s argument that “judgment” has been eliminated from the claim. However, even if the claim excluded “judgment,” the elimination of judgment per se from a claim is not the test for determining whether a claim recites a mental process. See MPEP 2106.04(a)(1)(III):
Accordingly, the "mental processes" abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, judgments, and opinions.
The courts do not distinguish between mental processes that are performed entirely in the human mind and mental processes that require a human to use a physical aid
. . .
Nor do the courts distinguish between claims that recite mental processes performed by humans and claims that recite mental processes performed on a computer.
(emphasis added). As such, Applicant’s argument is not persuasive.
Applicant further argues
The claimed combination of unique features of claim 1 results in several improvements over the state of the art as explained, for example, in paragraphs [0021] and [0030-0032]. Specifically, the interaction of users with generative artificial intelligence modules/systems/models (such as an LLM) is improved and provides a higher output quality which better fits the purpose of the application (user request).
Applicant’s paragraph 21 states the following:
Suitably combining an AAIM with a GAIM as suggested herein, allows the strengths of both modules to be combined: for specific analyzing tasks, a (specialized) AAIM can be used, which is much more efficient than the GAIM for these tasks, and the GAIM can be used where capabilities are required that the analytical models cannot provide, at least not in the desired quality. Technically, this is achieved by typically dynamically integrating/injecting the results of the AAIM into prompts to control the GAIM. As a result, an (energy) efficient, reliable, transparent and user-friendly AI system can be provided.
The determining steps for the AAIM, GAIM, and prompting module as recited in amended claim 1 can be all be practically performed in the human mind. Thus, they alone cannot be the basis for an improvement in technology. See MPEP 2106.05. (“. . it is important to keep in mind that an improvement in the abstract idea itself . . . is not an improvement in technology.”).
Examiner finds the additional elements do not reflect an improvement in technology disclosed in the specification or otherwise. That is, (1) the additional elements recite an outcome or solution (i.e. “configuring” the AAIM, module, and GAIM to perform the determining steps); without reciting how the outcome or solution is accomplished(i.e. how the AAIM, module, and GAIM are configured to perform the determining steps) and (2) the additional elements are recited at a high level of generality. See MPEP 2106.05(f) :
When determining whether a claim simply recites a judicial exception with the words "apply it" (or an equivalent), such as mere instructions to implement an abstract idea on a computer, examiners may consider the following:
(1) Whether the claim recites only the idea of a solution or outcome i.e., the claim fails to recite details of how a solution to a problem is accomplished.
. . .
(3) The particularity or generality of the application of the judicial exception.
See also MPEP 2106.05(a):
An important consideration in determining whether a claim improves technology is the extent to which the claim covers a particular solution to a problem or a particular way to achieve a desired outcome, as opposed to merely claiming the idea of a solution or outcome.
. . .
In this respect, the improvement consideration overlaps with other considerations, specifically the particular machine consideration (see MPEP § 2106.05(b)), and the mere instructions to apply an exception consideration (see MPEP § 2106.05(f)). Thus, evaluation of those other considerations may assist examiners in making a determination of whether a claim satisfies the improvement consideration.
(emphasis added). The additional limitations, at best, generally link the abstract idea to the field of use of large language models and computer programming. See MPEP 2106.05(h).
Finally, the additional elements “‘[a]dd nothing … that is not already present when the steps are considered separately’”. MPEP 2106.05 (I)(B)(quoting Alice). Thus, the additional elements, even when considered in combination in with the abstract idea elements, do not integrate the exception or recite an inventive concept.
Paragraph 30 states the following:
[0030]According to an embodiment, which can be combined with other embodiments described herein, determining metadata to be used for determining the prompt includes determining, based on the determined metadata, in particular on weights of the metadata, if further input data are required, and, if so, sending a further input data request to the interface, receiving the further input data via or from the interface, and determining the prompt based on the metadata, further metadata determined for the received further input data, and at least one of the received input data and the received further input data.
This language mirrors the claim language and thus sheds no light on whether the claim improves technology.
Paragraph 31 states:
[0031]Accordingly, the interaction of users with generative artificial intelligence modules/systems/models (such as an LLM) can be improved and/or can be explained and/or a more specific output and/or an output that better meets the expectations of the users (higher output quality) can be achieved. In particular, the interaction of users with the generative artificial intelligence modules/systems/models can be made more reliable. This is because the user initiated processes can be controlled by verified data from a transparent stream of input data instead of using in-transparent data from a hidden layer of the LLM / GAIM).
Examiner finds this paragraph merely asserts improvement(s) without the detail necessary to be apparent to one skilled in the art and thus does not improve technology. See MPEP 2106.04(d)(1).
In short, first the specification should be evaluated to determine if the disclosure provides sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The specification need not explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. Second, if the specification sets forth an improvement in technology, the claim must be evaluated to ensure that the claim itself reflects the disclosed improvement. That is, the claim includes the components or steps of the invention that provide the improvement described in the specification. The claim itself does not need to explicitly recite the improvement described in the specification (e.g., "thereby increasing the bandwidth of the channel
Additionally it is not clear which additional elements in the claimed invention reflect the teachings in paragraph 31.
Thus, Applicant’s argument that the claimed invention improves technology is not persuasive.
Applicant further argues
Moreover, generating the prompt based on metadata and their weights (weighted metadata) has been found to improve the efficiency of the system combining analytical artificial intelligence and generative artificial intelligence.
The AAI module, GAI module, and prompting module in the claimed invention do not provide significantly more because they generally link the abstract idea recited in the claim to the field of use of computer programming and LLMs and/or recite mere instructions to apply the exception. See MPEP 2106.05(f),(h). The determining steps alone cannot provide a basis for improving technology. See MPEP 2106.05. (“. . it is important to keep in mind that an improvement in the abstract idea itself . . . is not an improvement in technology.”). The additional elements “‘[a]dd nothing … that is not already present when the steps are considered separately’”. MPEP 2106.05 (I)(B)(quoting Alice). Thus, the claimed invention when considered as individual elements and as a whole do not integrate the exception and do not recite an inventive concept.
Applicant further argues “In particular, due to taking into account weights of the metadata for generating the prompt only the most relevant metadata is selected (see also claim 8). . . “
Taking into account weights of the metadata for generating the prompt can be practically performed in the human mind using evaluation and/or judgment. It cannot be the basis for an improvement to technology. See MPEP 2106.05. (“. . it is important to keep in mind that an improvement in the abstract idea itself . . . is not an improvement in technology.”).
Applicant further argues “and used by the generative artificial intelligence module in generating the output, thereby increasing the output quality at reduced numerical effort / lower energy consumption (with less relevant metadata not considered by the generative artificial intelligence module), see also paragraphs [0020, 21] and [00158].”
The limitations corresponding to “used by the generative artificial intelligence module in generating the output” are additional elements that do not provide significantly more than the abstract idea. They, for example, do not recite how the output is generated. See MPEP 2106.05(f). They also generally link the abstract idea to the field of use of LLMs/generative AI/computer programming. See MPEP 2106.05(h). As such, they additional limitations do not provide significantly more than the abstract idea. Applicant’s argument is therefore not persuasive.
Applicant further argues
In particular, due to taking into account weights of the metadata for generating the prompt, key performance indicators, patterns, trends and anomaly signals in the meta-data can be considered (see claim 8) and used by the generative artificial intelligence module in generating the output, thereby increasing the output precision resulting in only adding this specific data to the prompt.
Taking into account weights of the metadata to create a prompt can be practically performed in the human mind. Considering “key performance indicators, patterns, trends and anomaly signals” can be practically performed in the human mind using evaluation and judgment. Using a GAIM module (i.e. a conventional LLM) to perform these mental tasks amounts to “apply it” using a conventional LLM and/or generally linking the abstract idea to the field of use of LLMs and/or generative AI. The claimed invention does not improve the LLM models themselves. See Recentive Analytics, Inc. v. Fox Corp., No. 2023-2437 (Fed. Cir. Apr. 18, 2025):
Machine learning is a burgeoning and increasingly important field and may lead to patent-eligible improvements in technology. Today, we hold only that patents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under § 101.
(emphasis added).
Applicant further argues
As a result, specific details in the data are given greater weight than the generative Al model's knowledge, leading to more precise results. In the result, large/huge amounts of data can be processed very efficiently to provide higher quality outputs.
Giving greater weight to metadata is a task that can be practically performed in the human mind. It merely requires evaluation and judgment of the metadata. Mental tasks cannot form the basis for improvement in technology. See MPEP 2106.05. (“. . it is important to keep in mind that an improvement in the abstract idea itself . . . is not an improvement in technology.”). This argument is therefore not persuasive.
Applicant argues
Applicants respectfully submit that Shen is completely silent and makes no teaching or suggest with respect to determining weights of the metadata and using the weights of the metadata for determining the prompt, defined by amended independent claim 1.
This argument is not persuasive. Shen teaches determining weights of metadata at least in para. 31, 39, 49, and 55. See also Non-Final rejection 1/27/2026 at pages 45-46 and page 50. The facets in Shen effectively weigh certain metadata content higher than others because it “forces the machine learning model to only rely upon relevant data and thereby prevents hallucination.” See Shen para. 55. This process is used to generate prompts. For example, products that a user is subscribed to has a higher weight than products that the user is not subscribed to. See Shen at para. 81 (“In some embodiments, the processing device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to.”).
Applicant further argues
The Examiner finds that metadata used in chat completion step 275 in Fig. 2 of Shen, which is reproduced below, include "facets".
FIG.2 /*
Fig. 3 of Shen is also reproduced below.
FIG. 3 However, Shen fails to teach or suggest anything with respect to determining weights of the metadata and using the weights of the metadata for determining the prompt, as defined by amended claim 1.
This argument is not persuasive. Shen teaches determining weights of metadata at least in para. 31, 39, 49, and 55. See also Non-Final rejection 1/27/2026 at pages 45-46 and page 50. The facets in Shen effectively weigh certain content higher than others because they “forces the machine learning model to only rely upon relevant data and thereby prevents hallucination.” See Shen para. 55. This differently (i.e. higher) weighted metadata is used to generate prompts. See Shen at para. 81 (“In some embodiments, the processing device creates the standardized prompt based on the metadata from the user input. For example, facet-based response generation component 150 uses a different prompt based on the products that the user of user system 110 is subscribed to.”).
Applicant further argues
As described above, the claimed combination of features of claim 1 provides several advantages over the state of the art, such as described by at least paragraphs [0021], [0030] to [0032] of the originally filed application. Shen fails to provide such advantages.
It is improper to import limitations from the specification into the claims. See In re Van Geuns, 988 F.2d 1181, 26 USPQ2d 1057 (Fed. Cir. 1993). Shen is not required to teach unrecited “advantages” disclosed in the specification in order to anticipate the claim. As such, Applicant’s argument that Shen fails to anticipate the claim because it fails to disclose unclaimed “advantages” is not persuasive.
Conclusion
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 ALBERT M PHILLIPS, III whose telephone number is (571)270-3256. The examiner can normally be reached 10a-6:30pm EST M-F.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Ann J Lo can be reached at (571) 272-9767. 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.
/ALBERT M PHILLIPS, III/Primary Examiner, Art Unit 2159
Prosecution Insights