DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
Claims 1-25 were previously pending in this application. The amendment filed 30 June 2026 has been entered and the following has occurred: Claims 1, 6, 14, 20-21, & 25 have been amended. Claims 4 & 13 have been cancelled. Claims 26-27 have been added.
Claims 1-3, 5-12, & 14-27 remain pending in the application
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 1-3, 5-12, & 14-27 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., a law of nature, a natural phenomenon, or an abstract idea) without significantly more.
The claims recite subject matter within a statutory category as a process (claims 6-12, 14-24), machine (claims 1-2, 4-5, & 26-27), and manufacture (claim 25) (Subject Matter Eligibility (SME) Test Step 1: Yes) which recite steps of:
receiving, by at least one processor via a network interface, an electronic data package from a source node;
obtaining a document image derived from the electronic data package;
generating extracted text and location data from the document image, wherein the location data defining mapped locations of the extracted text within the document image, the location data comprises a respective structural metadata identifier and a respective mapped location assigned to each of a plurality of text chunks of the extracted text, and
each structural metadata identifier is mapped to geometric coordinates defining the respective mapped location of the corresponding text chunk within the document image;
generating, using a machine learning (ML) component, a structured response by:
generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node, wherein the generating the prompt payload comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text;
providing the prompt payload to a generative AI model of the ML component; and
receiving the structured response from the generative AI model, wherein the structured response comprises:
(i) at least one compatibility indicator determined based on the capability specification of the service providing node, and
(ii) a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator, wherein the source citation references at least one specific structural metadata identifier assigned to the at least one relevant text chunk;
determining whether the at least one compatibility indicator satisfies the node capability specification of the service providing node; and
generating a verification interface for presentation, wherein the verification interface comprises a visual overlay that visually associates, based on at least one of the geometric coordinates mapped to the specific structural metadata identifier referenced by the source citation or the mapped location of the at least one relevant text chunk, the at least one compatibility indicator with the at least one relevant text chunk on the document image, wherein the verification interface visually distinguishes content generated by the generative AI model from content directly extracted from the electronic data package.
These steps of obtaining a document image, generating extracted text and location data from the document image, generating a structured response, generating a prompt payload and providing said payload to a generative AI model, receiving a structured response from the AI model, determining whether the at least one compatibility indicator satisfies the node capability specification, and generating a verification interface for presentation to the user, as drafted, under the broadest reasonable interpretation, includes performance of the limitation in the mind but for recitation of generic computer components. That is, other than reciting steps as performed by the generic computer components, nothing in the claim element precludes the step from practically being performed in the mind. For example, but for the obtaining a document image and generating extracted text and location data from the document image language, generating extracted text and location data in the context of this claim encompasses a mental process of a person/user determining certain textual and/or picture data from a received document. Similarly, the limitation of generating a structured response, as drafted, is a process that, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, such as a person or user determining a response to a received document. For example, but for the generating a prompt payload to be provided to a generative AI model of the ML component and receiving a structured response in from the model language, generating a prompt payload and receiving a response in the context of this claim encompasses a mental process of the user determining prompts to provide to a computer or computerized model, but recited for generative AI, in particular, and receiving an associated output from using the computerized model as a tool for analysis. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation in the mind but for the recitation of generic computer components, then it falls within the “Mental Processes” grouping of abstract ideas. Accordingly, the claim recites an abstract idea.
Dependent claims recite additional subject matter which further narrows or defines the abstract idea embodied in the claims (such as claims 2-3, 5, 7-12, 14-24 & 26-27, reciting particular aspects of how generating prompts/structured responses, model validation/verification, and/or logging various historical actions may be performed in the mind but for recitation of generic computer components) (SME Test Step 2A, Prong 1: Yes).
This judicial exception is not integrated into a practical application. In particular, the additional elements do not integrate the abstract idea into a practical application, other than the abstract idea per se, because the additional elements amount to no more than limitations which:
amount to mere instructions to apply an exception (such as recitation of a network interface, an electronic data package, a source node, a storage device, computer-executable instructions, a processor, a machine learning component, a generative AI model of the ML component, a verification interface, amounts to invoking computers as a tool to perform the abstract idea, see Applicant’s Specification [0078]-[0079] for a network interface; Spec [0102] for an electronic data package; Spec [0065] for a source node; Spec [0276] for a storage device; [0869] for computer-executable instructions; Spec [0842] for a processor; Spec [0652] for a machine learning component; Spec [0277] for a generative AI model of the ML component; Spec [0096] for a verification interface; see MPEP 2106.05(f));
add insignificant extra-solution activity to the abstract idea (such as recitation of obtaining an electronic data package and/or document image derived from said package, receiving a structured response from the generative AI model, providing the prompt payload to a generative AI model, and receiving the structured response from the generative AI model amounts to mere data gathering; recitation of generating extracted text and location data from the document image, generating a prompt payload comprising extracted text and node capability specification, determining whether the at least one compatibility indicator satisfies the node capability specification of the service providing node, the location data defining mapped locations of the extracted text within the document image, the location data comprising a respective structural metadata identifier and a respective mapped location assigned to each of a plurality of text chunks of the extracted text amounts to selecting a particular data source or type of data to be manipulated; recitation of generating a verification interface for presentation, generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node, wherein the generating the prompt payload comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text amounts to insignificant application, see MPEP 2106.05(g); recitation of generating a structured response comprising a compatibility indicator and a source citation and/or generating a verification interface for presentation amounts to gathering and analyzing information using conventional techniques and displaying the result, TLI Communications, see MPEP 2106.05(a)(II));
generally link the abstract idea to a particular technological environment or field of use (such as recitation of a machine learning component and/or generative AI model of the ML component, see MPEP 2106.05(h)).
Dependent claims recite additional subject matter which amount to limitations consistent with the additional elements in the independent claims (such as claims 2-3, 5, 7-12, 14-24 & 26-27, which recite limitations relating to a generative AI model, a processor, a network interface, an electronic data package, a verification interface, an external electronic health record system, additional limitations which amount to invoking computers as a tool to perform the abstract idea, see Applicant’s Specification [0277] for a generative model; [0842] for a processor; [0078]-[0079] for a network interface; [0102] for an electronic data package; [0096] for a verification interface; [0155] for an external electronic health record, see MPEP 2106.05(f); claims 2-3, 10, 12, 14, 19-21, & 24, which recite limitations relating to specifying received data parameters, such as the node capability specification, the recommendation comprising a rejection of the service providing node, etc., receiving the electronic data package via the network interface, receiving location data comprising location tags assigned to discrete chunks of the extracted text and/or geometric coordinates for discrete elements of the extracted text, receiving the structured response containing content extracted from the electronic data package and AI-generated content, receiving an electronic data package comprising one or more files or data streams, receiving the structured response, transmitting the structured response to an external EHR system, additional limitations which add insignificant extra-solution activity to the abstract idea which amounts to mere data gathering; claims 2, 5, 7, 12, 15, & 20-21, which recite limitations relating to defining a clinical question and set of valid answer options, embedding the location tags within the extracted text, mapping specific location tags to geometric coordinates for the visual overlay, generating the prompt payload by aggregating the plurality of clinical questions and extracted text into a single prompt payload, determining the compatibility, determining whether the electronic data package is incomplete based on varying data, additional limitations which add insignificant extra-solution activity to the abstract idea by selecting a particular data source or type of data to be manipulated; claims 5, 9, 16-18, 20, 23, & 26-27, which recite limitations relating to determining a recommendation regarding the service providing node, executing a data sanitization process on the extracted text, utilizing a generative AI model to process clinical descriptions of sensitive medical conditions without censorship, storing a prompt payload, the structured response, and the source citation in a log, assigning a confidence score to the at least one compatibility indicator, additional limitations which amount to insignificant application; claims 2-3, 5, 7-12, & 14-24, which generally relate the limitations to AI-generated content, a machine learning, certain file/data stream, and/or forward-facing interface environments/processes, additional limitations which generally link the abstract idea to a particular technological environment or field of use; claims 3, 5, 8-11, 14, & 22-23, & 26-27, which recite limitations relating to generating one or more interfaces for the analyses performed for output to a user, or certain visual elements to be found therein amounts to gathering and analyzing information using conventional techniques and displaying the result, TLI Communications, see MPEP 2106.05(a)(II)). Looking at the limitations as an ordered combination adds nothing that is not already present when looking at the elements taken individually. There is no indication that the combination of elements improves the functioning of a computer or improves any other technology. Their collective functions merely provide conventional computer implementation and do not impose a meaningful limit to integrate the abstract idea into a practical application (SME Test Step 2A, Prong 2: No).
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to discussion of integration of the abstract idea into a practical application, the additional elements amount to no more than mere instructions to apply an exception, add insignificant extra-solution activity to the abstract idea, and generally link the abstract idea to a particular technological environment or field of use. Additionally, the additional limitations, other than the abstract idea per se, amount to no more than limitations which:
amount to elements that have been recognized as well-understood, routine, and conventional activity in particular fields (such as obtaining an electronic data package and/or document image derived from said package, receiving a structured response from the generative AI model, providing the prompt payload to a generative AI model, and receiving the structured response from the generative AI model, e.g., receiving or transmitting data over a network, Symantec, MPEP 2106.05(d)(II)(i); generating extracted text and location data from the document image, generating a prompt payload comprising extracted text and node capability specification, determining whether the at least one compatibility indicator satisfies the node capability specification of the service providing node, generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node, wherein the generating the prompt payload comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text e.g., performing repetitive calculations, Flook, MPEP 2106.05(d)(II)(ii); maintaining one or more ML components/generative AI models and constraints thereof, maintaining one or more compatibility indicators, maintaining one or more citations that are relevant from extracted text, e.g., electronic recordkeeping, Alice Corp., MPEP 2106.05(d)(II)(iii); storing computerized instructions for performance of the steps recited, storing computerized instructions for performance of the ML component/generative AI model, storing one or more received prompt payloads, storing generated information/outputs from the models, e.g., storing and retrieving information in memory, Versata Dev. Group, MPEP 2106.05(d)(II)(iv); generating extracted text and location data from a document image/electronic data package, e.g., electronic scanning or extracting data from a physical document, Content Extraction, MPEP 2106.05(d)(II)(v); utilizing the substantially computerized methods described, such as via an interface, and/or generating an interface for presentation to the user etc., e.g., a web browser’s back and forward button functionality, Internet Patent Corp., MPEP 2106.05(d)(II)(ii)).
Dependent claims recite additional subject matter which, as discussed above with respect to integration of the abstract idea into a practical application, amount to invoking computers as a tool to perform the abstract idea. Dependent claims recite additional subject matter which amount to limitations consistent with the additional elements in the independent claims (such as claims 2-3, 5, 7-12, 14-24 & 26-27, additional limitations which amount to elements that have been recognized as well-understood, routine, and conventional activity in particular fields: claims 2-3, 14, 19-21, 24, & 26-27, which recite limitations relating to specifying received data parameters, such as the node capability specification, receiving the electronic data package via the network interface, receiving the structured response containing content extracted from the electronic data package and AI-generated content, receiving an electronic data package comprising one or more files or data streams, receiving the structured response, transmitting the structured response to an external EHR system, e.g., receiving or transmitting data over a network, Symantec, MPEP 2106.05(d)(II)(i); claims 5, 9, 12, 15-16, 20, 23, & 26-27, which recite limitations relating to determining a recommendation regarding the service providing node, executing a data sanitization process on the extracted text, assigning a confidence score to the at least one compatibility indicator, e.g., performing repetitive calculations, Flook, MPEP 2106.05(d)(II)(ii); claim 17-18, which recite limitations relating to relating to maintaining permissions for access to one or more datastores, source nodes, etc., storing a prompt payload, the structured response, and the source citation in a log, e.g., electronic recordkeeping, Alice Corp., MPEP 2106.05(d)(II)(iii); claims 2-3, 5, 7-12, & 14-24, which recite limitations relating to storing a prompt payload, the structured response, and the source citation in a log, e.g., storing and retrieving information in memory, Versata Dev. Group, MPEP 2106.05(d)(II)(iv); claims 2-3, 5, 7-12, & 14-24, which all generally recite limitations relating to or specifying the manner of utilizing a machine learning component and/or generative AI model to extract text and location data from a document, e.g., electronic scanning or extracting data from a physical document, Content Extraction, MPEP 2106.05(d)(II)(v); claim 22, which recites limitations relating to receiving user interaction to determine when to render a visual overlay, e.g., a web browser’s back and forward button functionality, Internet Patent Corp., MPEP 2106.05(d)(II)(ii); claim 17, utilizing a generative AI model to process clinical descriptions of sensitive medical conditions without censorship, e.g., see Zaidi Par [0057] & [0074]; Nam Par [0072] which discloses the well-understood, routine, and/or conventional nature of utilizing a generative AI model for purposes of processing clinical descriptions and/or generating medically relevant content). Looking at the limitations as an ordered combination adds nothing that is not already present when looking at the elements taken individually. There is no indication that the combination of elements improves the functioning of a computer or improves any other technology. Their collective functions merely provide conventional computer implementation (SME Test Step 2B: No).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 3, 5-6, 8-9, 14-15, 18, 21-23, & 25-27 are rejected under 35 U.S.C. 103 as being unpatentable over Couleaud et al. (U.S. Patent Publication No. 2025/0078347), hereinafter “Couleaud” in view of Wang et al. (U.S. Patent Publication No. 2026/0030861), hereinafter “Wang”.
Claim 1 –
Regarding Claim 1, Couleaud discloses a system, comprising:
a network interface configured to receive an electronic data package from a source node (a “node” is understood simply constitute any source or point of origination from which data/media originates, therefore see Couleaud Par [0036], [0085], [0094], & [0097] which discloses receiving electronic data from one or more sources and accessing/received said data over a network, such as over any suitable data interface or remote server implementation, thereby constituting a network interface);
at least one storage device storing computer-executable instructions (See Couleaud Par [0095] & [0099] which discloses one or more storage devices, i.e. memories, containing computerized instructions therein); and
at least one processor coupled to the at least one storage device and configured to execute the computer-executable instructions to perform operations (See Couleaud Par [0095] & [0099] which discloses one or more storage devices, i.e. memories, containing computerized instructions therein for performance of steps by a computer device or other control circuitry) comprising:
receiving, by the at least one processor via a network interface, an electronic data package from a source node (See Couleaud Par [0050] & Fig. 1, el. 102, 110 which discloses receiving textual prompt and image associated with said textual prompt);
obtaining a document image derived from the electronic data package (See Couleaud Par [0050] & Fig. 1, el. 102, 110 which discloses receiving textual prompt and image associated with said textual prompt; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image);
generating extracted text and location data of the document image (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]), wherein
generating extracted text and location data from the document image, wherein the location data defining mapped locations of the extracted text within the document image, the location data comprises a respective structural metadata identifier and a respective mapped location assigned to each of a plurality of text chunks of the extracted text (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values), and
each structural metadata identifier is mapped to geometric coordinates defining the respective mapped location of the corresponding text chunk within the document image (while not “geometric coordinates”, see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values);
generating, using a machine learning (ML) component, a structured response by:
generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node (See Couleaud Par [0050] & Fig. 1 which discloses generating a set of textual prompts, i.e. understood to constitute a “payload”, corresponding to a received textual prompt and image associated with said textual prompt; See Couleaud Par [0077] which discloses “node capability specifications” at least by each layer may comprise one or more nodes that may be associated with learned parameters (e.g., weights and/or biases), and/or connections between nodes may represent parameters (e.g., weights and/or biases) learned during training (e.g., using backpropagation techniques, and/or any other suitable techniques), and in some embodiments, the nature of the connections may enable or inhibit certain nodes of the network, constituting node capabilities), wherein the generating the prompt payload comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]; See Couleaud Par [0002] which specifically mentions “embedding” given text caption);
providing the prompt payload to a generative AI model of the ML component (See Couleaud Par [0002] which discloses the text-to-photo synthesis model using a combination of a text encoder and a generative neural network to generate images from textual descriptions; See Couleaud Par [0008] which discloses leveraging generative AI models to guide the generation of images from text prompts; See Couleaud Par [0050] & Fig. 1 which discloses generating a set of textual prompts, i.e. understood to constitute a “payload”, corresponding to a received textual prompt and image associated with said textual prompt and that payload being provided to the machine learning model (comprising the generative AI models to guide the generation of images from said text prompts of Couleaud Par [0008])); and
receiving the structured response from the generative AI model (See Couleaud Par [0074] which discloses the image processing system may input data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved), wherein the structured response comprises:
(i) at least one compatibility indicator determined based on the capability specification of the service providing node (See Couleaud Par [0074] which discloses the image processing system may data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator), and
wherein the source citation references at least one specific structural metadata identifier assigned to the at least one relevant text chunk (see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values);
determining whether the at least one compatibility indicator satisfies the node capability specification of the service providing node (See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute determining whether the compatibility indicator satisfies the specification for the client/entity); and
generating a verification interface for the source node, wherein the verification interface comprises a visual overlay that visually associates, based on at least one of the geometric coordinates mapped to the specific structural metadata identifier referenced by the source citation or the mapped location of the at least one relevant text chunk, the at least one compatibility indicator with the at least one relevant text chunk on the document image (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein and/or metadata identifiers, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values), wherein the verification interface visually distinguishes content generated by the generative AI model from content directly extracted from the electronic data package (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance, i.e. confidence, level is achieved, such as a 95% accuracy level, such that annotated indications of the correct or desired outputs for given input(s) would therefore be distinguished from elements that fall below said confidence score or incorrect outputs, because there is a lack of annotated indication).
While Couleaud generally discloses generating a compatibility indicator for one or more text chunks extracted from an image, Couleaud does not seem to disclose an associated source citation identifying at least one of the relevant text chunks as required by the following limitation:
(ii) a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator,
However, Wang discloses a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator (See Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item; See Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051]; See Wang Par [0131] which discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output). The disclosure of Wang is directly applicable to the disclosure of Couleaud because both disclosures share limitations and capabilities, such as being directed toward automatically generating prompts for performance of generative AI operations.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the disclosure of Couleaud which already discloses generating a compatibility indicator for one or more text chunks extracted from an image to further include an associated source citation identifying at least one of the relevant text chunk, as disclosed by Wang, because this allows for identifying associations between units of prompt and units of content of media item, such as to generate probabilities that a particular unit of a media item is associated with an individual object based on said established text and media features established (See Wang Par [0049]-[0051] & [0055]).
Claim 3 –
Regarding Claim 3, Couleaud and Wang disclose the system of claim 1 in its entirety. Couleaud further discloses a system, wherein:
the at least one processor is configured to perform the generating the verification interface using a serverless function trigger activated based on receipt of the electronic data package via the network interface (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Fig. 1 which discloses & [0074] which discloses an iterative process that is automatically performed until results stop improving or a certain performance level is achieved, i.e. understood to constitute a serverless function trigger since the iterations are performed automatically until a certain performance level is achieved).
Claim 5 –
Regarding Claim 5, Couleaud and Wang disclose the system of claim 1 in its entirety. Couleaud further discloses a system, wherein:
the determining whether the at least one compatibility indicator satisfies the node capability specification comprises determining a recommendation regarding the service providing node (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0081] which discloses outputting a recommendation to a user based on preference or historical user interactions, thereby regarding the service node); and
the verification interface is configured to display the recommendation to a reviewer and receive an input in response to the recommendation (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0081]-[0083] which discloses outputting a recommendation to a user via user interface or GUI based on preference or historical user interactions, thereby regarding the service node, such that this allows a user such platform control over the generated recommendation/content based on user input).
Claim 6 –
Regarding Claim 6, Couleaud discloses a method, comprising:
receiving, by the at least one processor via a network interface, an electronic data package from a source node (a “node” is understood simply constitute any source or point of origination from which data/media originates, therefore see Couleaud Par [0036], [0085], [0094], & [0097] which discloses receiving electronic data from one or more sources and accessing/received said data over a network, such as over any suitable data interface or remote server implementation, thereby constituting a network interface; See Couleaud Par [0050] & Fig. 1, el. 102, 110 which discloses receiving textual prompt and image associated with said textual prompt);
obtaining a document image derived from the electronic data package (See Couleaud Par [0050] & Fig. 1, el. 102, 110 which discloses receiving textual prompt and image associated with said textual prompt; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image);
generating extracted text and location data of the document image (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]);
generating extracted text and location data from the document image, wherein the location data defining mapped locations of the extracted text within the document image, the location data comprises a respective structural metadata identifier and a respective mapped location assigned to each of a plurality of text chunks of the extracted text (see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values), and
each structural metadata identifier is mapped to geometric coordinates defining the respective mapped location of the corresponding text chunk within the document image (while not “geometric coordinates”, see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values);
generating, using a machine learning (ML) component, a structured response by:
generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node (See Couleaud Par [0050] & Fig. 1 which discloses generating a set of textual prompts, i.e. understood to constitute a “payload”, corresponding to a received textual prompt and image associated with said textual prompt; See Couleaud Par [0077] which discloses “node capability specifications” at least by each layer may comprise one or more nodes that may be associated with learned parameters (e.g., weights and/or biases), and/or connections between nodes may represent parameters (e.g., weights and/or biases) learned during training (e.g., using backpropagation techniques, and/or any other suitable techniques), and in some embodiments, the nature of the connections may enable or inhibit certain nodes of the network, constituting node capabilities);
wherein the generating the prompt payload comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]; See Couleaud Par [0002] which specifically mentions “embedding” given text caption).
providing the prompt payload to a generative AI model of the ML component (See Couleaud Par [0002] which discloses the text-to-photo synthesis model using a combination of a text encoder and a generative neural network to generate images from textual descriptions; See Couleaud Par [0008] which discloses leveraging generative AI models to guide the generation of images from text prompts; See Couleaud Par [0050] & Fig. 1 which discloses generating a set of textual prompts, i.e. understood to constitute a “payload”, corresponding to a received textual prompt and image associated with said textual prompt and that payload being provided to the machine learning model (comprising the generative AI models to guide the generation of images from said text prompts of Couleaud Par [0008])); and
receiving the structured response from the generative AI model (See Couleaud Par [0074] which discloses the image processing system may input data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved), wherein the structured response comprises:
(i) at least one compatibility indicator determined based on the capability specification of the service providing node (See Couleaud Par [0074] which discloses the image processing system may data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator), and
determining whether the at least one compatibility indicator satisfies the node capability specification of the service providing node (See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute determining whether the compatibility indicator satisfies the specification for the client/entity); and
generating a verification interface for the source node, wherein the verification interface comprises a visual overlay that visually associates, based on at least one of the geometric coordinates mapped to the specific structural metadata identifier referenced by the source citation or the mapped location of the at least one relevant text chunk, the at least one compatibility indicator with the at least one relevant text chunk on the document image (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator), wherein the verification interface visually distinguishes content generated by the generative AI model from content directly extracted from the electronic data package (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0052]-[0053] which discloses filling one or more holes or empty regions prior to inputting images to a subsequent machine learning model, such that the original image contains content directly from the extracted image, and the one or more filled holes contain AI-generated content, constituting distinguished content).
While Couleaud generally discloses generating a compatibility indicator for one or more text chunks extracted from an image, Couleaud does not seem to disclose an associated source citation identifying at least one of the relevant text chunks as required by the following limitation:
(ii) a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator;
However, Wang discloses a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator (See Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item; See Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051]; See Wang Par [0131] which discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the disclosure of Couleaud which already discloses generating a compatibility indicator for one or more text chunks extracted from an image to further include an associated source citation identifying at least one of the relevant text chunk, as disclosed by Wang, because this allows for identifying associations between units of prompt and units of content of media item, such as to generate probabilities that a particular unit of a media item is associated with an individual object based on said established text and media features established (See Wang Par [0049]-[0051] & [0055]).
Claim 8 –
Regarding Claim 8, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
the visual overlay comprises a bounding box or a color highlight (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0008] & [0117] which discloses boundaries and/or edges and/or other characteristics, i.e. bounding boxes, of objects or portions are considered along with other attributes for recognition purposes; While not relied upon since Couleaud effectively discloses the “bounding box” visual overlay option, for purposes of advancing prosecution, see Wang Par [0031] which discloses highlighting pixels associated with specific objects).
Claim 9 –
Regarding Claim 9, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
the determining whether the at least one compatibility indicator satisfies the node capability specification comprises determining a recommendation regarding the service providing node (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0081] which discloses outputting a recommendation to a user based on preference or historical user interactions, thereby regarding the service node); and
the generating the verification interface comprises configuring the verification interface to display the recommendation to a reviewer and receive an input in response to the recommendation (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0081]-[0083] which discloses outputting a recommendation to a user via user interface or GUI based on preference or historical user interactions, thereby regarding the service node, such that this allows a user such platform control over the generated recommendation/content based on user input).
Claim 14 –
Regarding Claim 14, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
the structured response comprises content directly extracted from the electronic data package and AI-generated content (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0052]-[0053] which discloses filling one or more holes or empty regions prior to inputting images to a subsequent machine learning model, such that the original image contains content directly from the extracted image, and the one or more filled holes contain AI-generated content); and
the verification interface comprises a visual indication distinguishing the AI-generated content from the content directly extracted from the electronic data package and the content generated by the generative AI model (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0052]-[0053] which discloses filling one or more holes or empty regions prior to inputting images to a subsequent machine learning model, such that the original image contains content directly from the extracted image, and the one or more filled holes contain AI-generated content).
Claim 15 –
Regarding Claim 15, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
determining, based on the structured response, that the electronic data package is incomplete regarding a mandatory data element defined in the node capability specification (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0051] which discloses textual prompts being compared to or reconciled with text input to check whether any of textual prompts should be updated with information from text input, such that if said data does need to be updated, then the textual prompt is updated with information from text input that may be relevant to the corresponding extracted object); and
automatically triggering an electronic task to request additional information from the source node regarding the mandatory data element prior to determining whether the at least one compatibility indicator satisfies the node capability specification (ee Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0051] which discloses the system automatically performing textual prompts being compared to or reconciled with text input to check whether any of textual prompts should be updated with information from text input, such that if said data does need to be updated, then the textual prompt is updated with information from text input that may be relevant to the corresponding extracted object).
Claim 18 –
Regarding Claim 18, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud and Wang further disclose a method, wherein:
storing the prompt payload, the structured response, and the source citation in a log to generate an audit trail (“to generate an audit trail” is understood to be an intended result of a process step positively recited, and therefore is not given patentable weight, see MPEP 2111.04, See Couleaud Par [0075] which discloses input data, i.e. prompts, training data, and/or outputs, i.e. structured response, may be stored at any suitable device and/or server; See Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item; See Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051]; See Wang Par [0131] which discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output, i.e. source citation).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the disclosure of Couleaud which already discloses generating a compatibility indicator for one or more text chunks extracted from an image to further include an associated source citation identifying at least one of the relevant text chunk, as disclosed by Wang, because this allows for identifying associations between units of prompt and units of content of media item, such as to generate probabilities that a particular unit of a media item is associated with an individual object based on said established text and media features established (See Wang Par [0049]-[0051] & [0055]).
Claim 21 –
Regarding Claim 21, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
the generating the verification interface comprises using the location data to map the specific structural metadata identifier to the geometric coordinates to render the visual overlay (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation).
Claim 22 –
Regarding Claim 22, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
the visual overlay is dynamically rendered upon user interaction with the at least one compatibility indicator (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator, and/or the user interacting with said user interface to control various constraints of the model, and subsequently presenting an updated interface with the model’s performance with the modified constraints).
Claim 23 –
Regarding Claim 23, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud further discloses a method, wherein:
the deriving the at least one compatibility indicator comprises assigning a confidence score to the at least one compatibility indicator (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance, i.e. confidence, level is achieved, such as a 95% accuracy level), and
the verification interface visually distinguishes one of the at least one compatibility indicator that has a confidence score below a threshold (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance, i.e. confidence, level is achieved, such as a 95% accuracy level, such that annotated indications of the correct or desired outputs for given input(s) would therefore be distinguished from elements that fall below said confidence score or incorrect outputs, because there is a lack of annotated indication).
Claim 25 –
Regarding Claim 25, Couleaud discloses at least one non-transitory computer readable storage medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising (See Couleaud Par [0096] & Couleaud Par [0103]):
receiving, by the at least one processor via a network interface, an electronic data package from a source node (See Couleaud Par [0050] & Fig. 1, el. 102, 110 which discloses receiving textual prompt and image associated with said textual prompt);
obtaining a document image derived from the electronic data package (See Couleaud Par [0050] & Fig. 1, el. 102, 110 which discloses receiving textual prompt and image associated with said textual prompt; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image);
generating extracted text and location data of the document image, the location data defining mapped locations of the extracted text within the document image (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]);
generating extracted text and location data from the document image, wherein the location data defining mapped locations of the extracted text within the document image, the location data comprises a respective structural metadata identifier and a respective mapped location assigned to each of a plurality of text chunks of the extracted text (see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values), and
each structural metadata identifier is mapped to geometric coordinates defining the respective mapped location of the corresponding text chunk within the document image (while not “geometric coordinates”, see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values);
generating, using a machine learning (ML) component, a structured response by:
generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node (See Couleaud Par [0050] & Fig. 1 which discloses generating a set of textual prompts, i.e. understood to constitute a “payload”, corresponding to a received textual prompt and image associated with said textual prompt; See Couleaud Par [0077] which discloses “node capability specifications” at least by each layer may comprise one or more nodes that may be associated with learned parameters (e.g., weights and/or biases), and/or connections between nodes may represent parameters (e.g., weights and/or biases) learned during training (e.g., using backpropagation techniques, and/or any other suitable techniques), and in some embodiments, the nature of the connections may enable or inhibit certain nodes of the network, constituting node capabilities), wherein the generating the prompt payload comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text (See Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]; See Couleaud Par [0002] which specifically mentions “embedding” given text caption);
providing the prompt payload to a generative AI model of the ML component (See Couleaud Par [0002] which discloses the text-to-photo synthesis model using a combination of a text encoder and a generative neural network to generate images from textual descriptions; See Couleaud Par [0008] which discloses leveraging generative AI models to guide the generation of images from text prompts; See Couleaud Par [0050] & Fig. 1 which discloses generating a set of textual prompts, i.e. understood to constitute a “payload”, corresponding to a received textual prompt and image associated with said textual prompt and that payload being provided to the machine learning model (comprising the generative AI models to guide the generation of images from said text prompts of Couleaud Par [0008])); and
receiving the structured response from the generative AI model (See Couleaud Par [0074] which discloses the image processing system may input data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved), wherein the structured response comprises:
(i) at least one compatibility indicator determined based on the capability specification of the service providing node (See Couleaud Par [0074] which discloses the image processing system may data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator), and
wherein the source citation references at least one specific structural metadata identifier assigned to the at least one relevant text chunk (see Couleaud Par [0062] which discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image; See Couleaud Par [0078] & [0090] which discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values);
determining whether the at least one compatibility indicator satisfies the node capability specification of the service providing node (See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute determining whether the compatibility indicator satisfies the specification for the client/entity); and
generating a verification interface for the source node, wherein the verification interface comprises a visual overlay that visually associates, based on at least one of the geometric coordinates mapped to the specific structural metadata identifier referenced by the source citation or the mapped location of the at least one relevant text chunk, the at least one compatibility indicator with the at least one relevant text chunk on the document image (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator), wherein the verification interface visually distinguishes content generated by the generative AI model from content directly extracted from the electronic data package (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0044] which discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], and further discloses generating a map associating image and object with associated indicators of pixel values; See Couleaud Par [0052]-[0053] which discloses filling one or more holes or empty regions prior to inputting images to a subsequent machine learning model, such that the original image contains content directly from the extracted image, and the one or more filled holes contain AI-generated content, constituting distinguished content).
While Couleaud generally discloses generating a compatibility indicator for one or more text chunks extracted from an image, Couleaud does not seem to disclose an associated source citation identifying at least one of the relevant text chunks as required by the following limitation:
(ii) a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator;
However, Wang discloses a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator (See Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item; See Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051]; See Wang Par [0131] which discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the disclosure of Couleaud which already discloses generating a compatibility indicator for one or more text chunks extracted from an image to further include an associated source citation identifying at least one of the relevant text chunk, as disclosed by Wang, because this allows for identifying associations between units of prompt and units of content of media item, such as to generate probabilities that a particular unit of a media item is associated with an individual object based on said established text and media features established (See Wang Par [0049]-[0051] & [0055]).
While Couleaud generally discloses generating a compatibility indicator for one or more text chunks extracted from an image, Couleaud does not seem to disclose an associated source citation identifying at least one of the relevant text chunks as required by the following limitation:
(ii) a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator;
However, Wang discloses a source citation identifying at least one relevant text chunk from the extracted text that supports the at least one compatibility indicator (See Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item; See Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051]; See Wang Par [0131] which discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the disclosure of Couleaud which already discloses generating a compatibility indicator for one or more text chunks extracted from an image to further include an associated source citation identifying at least one of the relevant text chunk, as disclosed by Wang, because this allows for identifying associations between units of prompt and units of content of media item, such as to generate probabilities that a particular unit of a media item is associated with an individual object based on said established text and media features established (See Wang Par [0049]-[0051] & [0055]).
Claim 26 –
Regarding Claim 26, Couleaud and Wang discloses the system of Claim 1 in its entirety. Couleaud further discloses the system in its entirety:
the visual overlay is dynamically rendered upon user interaction with the at least one compatibility indicator (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator, and/or the user interacting with said user interface to control various constraints of the model, and subsequently presenting an updated interface with the model’s performance with the modified constraints).
Claim 27 –
Regarding Claim 27, Couleaud and Wang disclose the system of Claim 1 in its entirety. Couleaud further discloses the system in its entirety:
the visual overlay comprises at least one of a bounding box or a color highlight (See Couleaud Par [0063] which discloses a graphical user interface may be provided to allow control over each step of the layer generation techniques, enabling the user to be prompted to confirm or edit the original segmentation or refine prompts generated after segmentation; See Couleaud Par [0074] which discloses the image processing system may input training data into untrained model to generate outputs, such that respective outputs may be compared to ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator; See Couleaud Par [0008] & [0117] which discloses boundaries and/or edges and/or other characteristics, i.e. bounding boxes, of objects or portions are considered along with other attributes for recognition purposes; While not relied upon since Couleaud effectively discloses the “bounding box” visual overlay option, for purposes of advancing prosecution, see Wang Par [0031] which discloses highlighting pixels associated with specific objects).
Claims 2, 7, & 24 are rejected under 35 U.S.C. 103 as being unpatentable over Couleaud in view of Wang, further in view of Zaidi et al. (U.S. Patent Publication No. 2025/0095807), hereinafter “Zaidi”.
Claim 2 –
Regarding Claim 2, Couleaud and Wang disclose the system of claim 1 in its entirety. Couleaud and Wang do not further disclose a system, wherein:
the node capability specification defines a clinical question and a set of valid answer options for the clinical question; and
the generating the prompt payload comprises incorporating the set of valid answer options in the prompt payload to constrain the structured response of the generative AI model to a selected option from the set of valid answer options.
While Wang Par [0071] generally discloses one or more performed actions/applications of the generative AI system being generating an automated medical diagnostic determination (e.g., identifying one or more patient conditions/diseases based on a medical imaging MI), generating an automated patient wellbeing alarm (e.g., observing that a patient is in a dangerous state or condition at home, assisted living facility, medical inpatient facility, medical outpatient facility, etc.), and/or the like, Wang is generally silent on specifically defining a clinical question and a set of valid answer options for the clinical question.
However, Zaidi discloses a system, wherein the node capability specification defines a clinical question and a set of valid answer options for the clinical question (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question), and the generating the prompt payload comprises incorporating the set of valid answer options in the prompt payload to constrain the structured response of the generative AI model to a selected option from the set of valid answer options (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question, such that the response given by the generative model is further constrained into medically relevant or not being medically relevant depending on the medical question provided, i.e. generating results that include a set of filtered and classified facts such that non-medically relevant facts in the collection of facts can be discarded prior to generating a SOAP note; See Zaidi Par [0099] which discloses said automated prompt engineering designing prompts to artificial intelligence (AI), i.e. generative, models to produce/generate a more accurate/optimal response). The disclosure of Zaidi is directly applicable to the combined disclosure of Couleaud and Wang because the disclosures share limitations and capabilities, such as being directed towards automatically generating prompts for performance of generative AI operations.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang, which already discloses generating an automated medical diagnostic determination and/or generating an automated patient wellbeing alarm, to further specifically include incorporating a clinical question and a set of valid answer options in the prompt payload to constrain the structured response of the generative AI model to a selected option from the set of valid answer options, as disclosed by Zaidi, because this allows for non-medically relevant facts in the collection of facts being discarded prior to generating a SOAP note by the system (See Zaidi Par [0074]).
Claim 7 –
Regarding Claim 7, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud and Wang do not, but Zaidi does further disclose a method, wherein:
the node capability specification defines a clinical question and a set of valid answer options for the clinical question (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question), and
the generating the prompt payload comprises incorporating the set of valid answer options in the prompt payload to constrain the structured response of the generative AI model to a selected option from the set of valid answer options (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question, such that the response given by the generative model is further constrained into medically relevant or not being medically relevant depending on the medical question provided, i.e. generating results that include a set of filtered and classified facts such that non-medically relevant facts in the collection of facts can be discarded prior to generating a SOAP note; See Zaidi Par [0099] which discloses said automated prompt engineering designing prompts to artificial intelligence (AI), i.e. generative, models to produce/generate a more accurate/optimal response).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang, which already discloses generating an automated medical diagnostic determination and/or generating an automated patient wellbeing alarm, to further specifically include incorporating a clinical question and a set of valid answer options in the prompt payload to constrain the structured response of the generative AI model to a selected option from the set of valid answer options, as disclosed by Zaidi, because this allows for non-medically relevant facts in the collection of facts being discarded prior to generating a SOAP note by the system (See Zaidi Par [0074]).
Claim 24 –
Regarding Claim 24, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud and Wang do not disclose, but Zaidi further discloses a method, wherein:
transmitting the structured response to an external electronic health record (EHR) system (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question, such that the response given by the generative model is further constrained into medically relevant or not being medically relevant depending on the medical question provided, i.e. generating results that include a set of filtered and classified facts such that non-medically relevant facts in the collection of facts can be discarded prior to generating a SOAP note; See Zaidi Par [0010] & [0061] which discloses storing the generated SOAP note in the database, i.e. electronic health record associated with the patient, i.e. via transmission efforts via one or more platforms in communication).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang, which already discloses generating an automated medical diagnostic determination and/or generating an automated patient wellbeing alarm, to further specifically include transmitting the structured response to an external electronic health record (EHR) system, as disclosed by Zaidi, because this allows for automatically generating a SOAP note and providing the SOAP note to a digital assistant service, such that the SOAP note service can automatically record or document interactions between one or more entities, such as healthcare interactions between a provider and a patient (See Zaidi Par [0061]).
Claims 10-12 & 16-17 are rejected under 35 U.S.C. 103 as being unpatentable over Couleaud, in view of Wang, further in view of Neystadt et al. (U.S. Patent Publication No. 2026/0023842), hereinafter “Neystadt”.
Claim 10 –
Regarding Claim 10, Couleaud and Wang disclose the method of claim 9 in its entirety. Couleaud and Wang do not further disclose a method, wherein:
the recommendation comprises a rejection of the service providing node;
the method comprises causing, in the verification interface, the rejection to be displayed in association with a specific text chunk of the at least one text chunk as evidence for the rejection.
However, Neystadt discloses a method, wherein the recommendation comprises a rejection of the service providing node (See Neystadt Par [0111] which discloses denying requests, i.e. rejection, of requests or services from team-members that are not permitted by the organizational policy, but also discloses partial rejection of viewing certain portions of said data, such as via masking, modification, redaction, and/or replacement of one or more content-portions or data-portions); and the method comprises causing, in the verification interface, the rejection to be displayed in association with a specific text chunk of the at least one text chunk as evidence for the rejection (See Neystadt Par [0111] which discloses denying requests, i.e. rejection, of requests or services from team-members that are not permitted by the organizational policy, but also discloses partial rejection of viewing certain portions of said data, such as via masking, modification, redaction, and/or replacement of one or more content-portions or data-portions; See Neystadt Par [0093] which discloses displaying an “access denied” response to the user if end-user is blocked, i.e. rejected, from viewing requested data/query according to LLM constraints policy, such as based on inquiries/queries/prompts that are related to particular topics (e.g., sex, drugs, alcohol, tobacco) and/or that include particular words or keywords or strings (e.g., “gun” or “explosives” or “pornographic” or “cigarettes”), and/or that are expected to yield LLM-generated results that may include data or content that pertains to such topics and/or keywords). The disclosure of Neystadt is directly applicable to the combined disclosure of Couleaud and Wang, because the disclosures share limitations and capabilities, such as being directed towards generating content using generative AI models.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang which already discloses outputting a recommendation to a user via the service providing node, to further specifically include a rejection of the service providing node and for said rejection to be displayed, as disclosed by Neystadt, because this allows for enhancing and respecting security and ensuring that sensitive information is not accessible to unauthorized users (See Neystadt Par [0111]).
Claim 11 –
Regarding Claim 11, Couleaud, Wang, and Neystadt disclose the method of claim 10 in its entirety. Neystadt further discloses a method, wherein:
the causing the rejection to be displayed in association with the specific text chunk comprises causing at least one of the specific text chunk or a link to the specific text chunk to be displayed (See Neystadt Par [0111] which discloses denying requests, i.e. rejection, of requests or services from team-members that are not permitted by the organizational policy, but also discloses partial rejection of viewing certain portions of said data, such as via masking, modification, redaction, and/or replacement of one or more content-portions or data-portions; See Neystadt Par [0093] which discloses displaying an “access denied” response to the user if end-user is blocked, i.e. rejected, from viewing requested data/query according to LLM constraints policy, such as based on inquiries/queries/prompts that are related to particular topics (e.g., sex, drugs, alcohol, tobacco) and/or that include particular words or keywords or strings (e.g., “gun” or “explosives” or “pornographic” or “cigarettes”), and/or that are expected to yield LLM-generated results that may include data or content that pertains to such topics and/or keywords).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the combined disclosure of Couleaud, Wang, and Neystadt which already discloses outputting a recommendation to a user via the service providing node, to further specifically include a rejection of the service providing node and for said rejection to be displayed, including said specific text chunk, as disclosed by Neystadt, because this allows for enhancing and respecting security and ensuring that sensitive information is not accessible to unauthorized users by partial rejection of viewing certain portions of said data, such as via masking, modification, redaction, and/or replacement of one or more content-portions or data-portions (See Neystadt Par [0093] & [0111]).
Claim 12 –
Regarding Claim 12, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud and Wang do not disclose, but Neystadt further discloses a method, wherein:
determining that the at least one compatibility indicator indicates the service providing node is incompatible with the electronic data package (See Neystadt Par [0111] which discloses denying requests, i.e. rejection, of requests or services from team-members that are not permitted by the organizational policy, but also discloses partial rejection of viewing certain portions of said data, such as via masking, modification, redaction, and/or replacement of one or more content-portions or data-portions; See Neystadt Par [0093] which discloses displaying an “access denied” response to the user if end-user is blocked, i.e. rejected, from viewing requested data/query according to LLM constraints policy, such as based on inquiries/queries/prompts that are related to particular topics (e.g., sex, drugs, alcohol, tobacco) and/or that include particular words or keywords or strings (e.g., “gun” or “explosives” or “pornographic” or “cigarettes”), and/or that are expected to yield LLM-generated results that may include data or content that pertains to such topics and/or keywords);
identifying a second service providing node having a node capability specification that is compatible with the at least one compatibility indicator (See Neystadt Par [0096] which discloses may define an LLM access policy that would block or modify queries about “alcohol” or “drugs” if they are submitted by minors/students, but that would allow and/or would differently modify (e.g., would less modify) queries about those topics if they are submitted by a teacher/an adult); and
generating a second recommendation to route the electronic data package to the second service providing node (See Neystadt Par [0096] which discloses may define an LLM access policy that would block or modify queries about “alcohol” or “drugs” if they are submitted by minors/students, but that would allow and/or would differently modify (e.g., would less modify) queries about those topics if they are submitted by a teacher/an adult).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the combined disclosure of Couleaud and Wang, which already discloses determining and comparing outputs to a compatibility indicator, to further specifically include a determining that the at least one compatibility indicator indicates the service providing node is incompatible with the electronic data package, identifying a second service providing node having a node capability specification that is compatible with the at least one compatibility indicator, and generating a second recommendation to route the electronic data package to the second service providing node, as disclosed by Neystadt, because this allows for accounting user-specific, i.e. node specific, data or characteristics, in order to apply and/or enforce data access policy, thereby enhancing and respecting security and ensuring that sensitive information is not accessible to unauthorized users (See Neystadt Par [0096] & [0111]).
Claim 16 –
Regarding Claim 16, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud and Neystadt further disclose a method, wherein:
executing a data sanitization process on the extracted text by masking a predefined type of personal information in the electronic data package (See Couleaud Par [0044]-[0045] which discloses masking such that based on performing semantic segmentation of a text prompt, clipping masks may be perfromed to ensure consistency between text prompt and generated object images; See Neystadt Par [0017] which discloses a computerized method that are configured to constrain, limit, define, modify, and/or otherwise control the outputs that are generated by an LLM (or other AI-based tool) when organizational team-members query an organizational data lake; See Neystadt Par [0023] which discloses the pre-defined organizational Selective LLM Authorization Policy restricting or defining which particular users or types-of-users can or cannot access particular types of information (e.g., financial data, sales data, employee compensation data, password/credentials data), based on sensitivity, confidentiality constrains, legal constrains of the organization (such as HIPAA for health services or medical services providers, i.e. personal information), semantic classification and/or project relatedness, and may cause either blocking of the information or adaptation/adapting/modification of the information that is outputted to that querying user, and further describes masking or removing particular types of information based on said policy; See Neystadt Par [0048] which discloses sending outputs to Authorization Proxy Unit which checks whether or not the LLM-based output complies with the relevant rules of the pre-defined organizational Selective LLM Authorization Policy, and blocks or removes or deletes or masks one or more portions of the LLM-based output that do not comply with such relevant rules, and only then transfers the modified LLM-based output to the querying user; thereby providing dual-stage protection against access to sensitive information; See Neystadt Par [0132]-[0133] which specifically mentions a post-processing “sanitization” unit performing deleting or masking of content).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang which already discloses masking based on performing semantic segmentation of a text prompt, to further specifically include masking for a predefined type of personal information, as disclosed by Neystadt, because this provides dual-stage protection against access to sensitive information by blocking, removing, deleting, or masking one or more portions of the LLM-based output that do not comply with security/access policy, such as for HIPAA compliance (See Neystadt Par [0023] & [0048]).
Claim 17 –
Regarding Claim 17, Couleaud and Wang discloses the method of claim 6 in its entirety. Couleaud and Wang do not disclose, but Neystadt further discloses a method, wherein:
the generating the prompt payload comprises including a clinical authorization instruction in the prompt payload configured to override a default content moderation filter of the generative AI model, thereby enabling the generative AI model to process valid clinical descriptions of sensitive medical conditions without censorship (See Neystadt Par [0108]-[0110] which discloses real-time organizational context updates such that the system periodically (e.g., daily, weekly) updates the organizational context, reflecting changes in user roles, permissions, project membership, department membership, and data access events, ensuring that query responses remain accurate and relevant over time, such that adaptation of queries and responses are based on real-time user and data classifications, and therefore if a user’s security permissions were changed to be able to access certain content, the default moderation filter for said person would thereby be overridden, and such that prompts and queries are modified, i.e. the prompt payload is generated, based on said permissions).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang which already discloses masking based on performing semantic segmentation of a text prompt, to further specifically include override a default content moderation filter of the generative AI model, as disclosed by Neystadt, because this allows for dynamically adjusting the information presented to different users to enhance security and relevance (See Neystadt Par [0108]-[0110]).
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over Couleaud in view of Wang, in view of Toscano, III, et al. (U.S. Patent Publication No. 2024/0428908), hereinafter “Toscano”, further in view of Khashman et al. (U.S. Patent Publication No. 2025/0157602), hereinafter “Khashman”.
Claim 19 –
Regarding Claim 19, Couleaud and Wang disclose the method of claim 6 in its entirety. Wang further discloses a method, wherein:
the electronic data package comprises one or more files or data streams selected from the group consisting of: a rasterized image file (See Wang Par [0124] which discloses inputting and/or generating visualizations for viewing outputs of applications, such as rasterized images), an Extensible Markup Language (XML) file (See Wang Par [0130] which discloses the input comprising structured data, such as an XML file;), and a JavaScript Object Notation (JSON) packet (See Wang Par [0130] which discloses the input comprising structure data, such as a JSON file).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang which already discloses receiving one or more inputs, i.e. files or datastreams, to further include particular formats, such as a rasterized image file, XML file, or JSON packet, as further disclosed by Wang, because this allows for file types of various formats, both structued and/or unstructured, to be received and processed by the generative learning model (See Wang Par [0130]).
While Couleaud and Wang disclose receiving one or more inputs, i.e. files or datastreams, including a rasterized image file, XML file, or JSON packet, Couleaud and Wang do not explicitly disclose receiving inputs including a Portable Document Format (PDF) file, a Facsimile (Fax) image, and/or a Health Level Seven (HL7) message.
However, Toscano discloses receiving one or more inputs, i.e. files or datastreams, including a Portable Document Format (PDF) file and/or a Facsimile (Fax) image (See Toscano, III Par [0009]-[0010] & [0037] which discloses an intake process including a received data, such that a C-CDA document can be created that is in HL7 format, from one or more image files, such as TIFF (from a fax) or pdf with embedded text and extracts the text into a text file). The disclosure of Toscano is directly applicable to the disclosure of Couleaud and Wang, such as being directed towards automated algorithms, such as artificial intelligence, for generating content.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang, which already discloses receiving one or more inputs, i.e. files or datastreams, in a particular format such as a rasterized image file, XML file, and/or JSON packet, to further include file formats such as a PDF or Fax file, as disclosed by Toscano, because this allows for different modes of communication to be implemented into the content generation system for receiving input data, such as facsimile and/or email (See Toscano Par [0037]).
While Couleaud, Wang, and Toscano disclose receiving one or more inputs, i.e. files or datastreams, including a rasterized image file, XML file, or JSON packet, a Portable Document Format (PDF) file, and/or a Facsimile (Fax) image, and Toscano specifically mentions creating documents in an HL7 format, Couleaud, Wang, and Toscano do not explicitly disclose receiving one or more inputs, i.e. files or datastreams, including a Health Level Seven (HL7).
However, Khashman discloses receiving one or more inputs, i.e. files or datastreams, including a Health Level Seven (HL7) (See Khashman Par [0037]-[0039] & [0043] which discloses automated content extraction and generation based on clinical notes, such as by receiving and inputting one or more HL7 messages comprising clinical note data and/or other types of electronic health record (EHR) data to the data access system and subsequently to the NLP subsystem, such as for generating a clinical recommendation and/or registry associated therewith). The disclosure of Khashman is directly applicable to the combined disclosure of Couleaud, Wang, and Toscano, because the disclosures share limitations and capabilities, such as being directed towards automated algorithms for generation of content based on inputs received.
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud, Wang, and Khashman, which already discloses receiving one or more inputs, i.e. files or datastreams, in a particular format such as a rasterized image file, XML file, JSON packet, PDF, and/or Fax file to further include receiving one or more inputs, i.e. files or datastreams, including a Health Level Seven (HL7), as disclosed by Khashman, because this allows for a subscriber to periodically or in real-time send one or more HL7 messages comprising clinical note data and/or other types of electronic health record (EHR) data to the data access system for use by the content generation system (See Khashman Par [0038]-[0039])
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over Couleaud in view of Wang in view of Zaidi, further in view of Toscano.
Claim 20 –
Regarding Claim 20, Couleaud and Wang disclose the method of claim 6 in its entirety. Couleaud, Wang, and Zaidi further disclose a method, wherein:
the node capability specification defines a plurality of clinical questions (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question);
the generating the prompt payload comprises aggregating the plurality of clinical questions and the extracted text into a single prompt payload to execute a single inference pass by the generative AI model (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question; See Zaidi Fig. 8 Par [0059] which discloses iteratively refine/improving the prompt into a final production prompt); and
the receiving the structured response comprises receiving a single JavaScript Object Notation (JSON) packet (See Toscano Par [0042] & [0053]-[0055] which discloses outputting or creating a JSON formatted document by the AI/ML model, such as from the extracted content, and further specifically discloses at Toscano Par [0056] that AI is trained to provide a confidence score based for example on the number of errors it found to correct and upper and lower ranges of the converted data, such that ) containing, for each of the plurality of clinical questions:
(i) a specific compatibility indicator of the at least one compatibility indicator (See Couleaud Par [0074] which discloses the image processing system may data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a compatibility indicator),
(ii) an answer determined based on a portion of the at least one relevant text chunk (See Zaidi Par [0057] which discloses consideration of information/inputs for generation of content, including questions from patients and relevant medical conditions or assessments, i.e. clinical questions; See Zaidi Par [0074] which discloses machine-learning model prompt for generating the set of filtered facts can include a query and context information, such as a query for classifying facts in the collection of facts and assigning labels to the classified facts, such as “Please classify and label each of the facts in the following list of facts as being medically relevant or not being medically relevant. Please generate a set of filtered facts based on the classification and labels”, i.e. a clinical question and a set of valid answer options/facts for the question; See Zaidi Fig. 8 Par [0059] which discloses iteratively refine/improving the prompt into a final production prompt), and
(iii) a portion of the source citation corresponding to the portion of the at least one relevant text chunk in support of the compatibility indicator (See Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item; See Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051]; See Wang Par [0131] which discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output, i.e. source citation).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the disclosure of Couleaud which already discloses generating a compatibility indicator for one or more text chunks extracted from an image to further include an associated source citation identifying at least one of the relevant text chunk, as disclosed by Wang, because this allows for identifying associations between units of prompt and units of content of media item, such as to generate probabilities that a particular unit of a media item is associated with an individual object based on said established text and media features established (See Wang Par [0049]-[0051] & [0055]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud and Wang to further specifically include incorporating a clinical question and a set of valid answer options in the prompt payload to constrain the structured response of the generative AI model to a selected option from the set of valid answer options, as disclosed by Zaidi, because this allows for non-medically relevant facts in the collection of facts being discarded prior to generating a SOAP note by the system (See Zaidi Par [0074]).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the combined disclosure of Couleaud, Wang, and Zaidi, which discloses the generation of a document containing various compatibility indicators, clinical questions and relevant answers, and/or source citations, to further include JSON formatting, as disclosed by Toscano, because 's a lightweight data interchange format that's easy for machines to parse and generate and can be easily used by other programs or AI systems for further processing, such as database storage, information retrieval, or integration into other applications (See Toscano Par [0053]-[0054]).
Response to Arguments
Applicant's arguments filed 30 June 2026 have been fully considered but they are not persuasive:
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 14-18 of Arguments/Remarks that the Office Action’s characterization of the limitations reciting a mental process do not survive the amendments. More specifically, Applicant argues that generating extracted text and location data cannot be performed by a person, such as “determining certain textual and/or picture data from a received document”, because it requires constructing a coordinate map that assigns to each of a plurality of text chunks a structural metadata identifier and corresponding geometric coordinates defining the chunk’s position within the document image, such that a human cannot assign geometric coordinates to text chunks, and therefore precludes the limitations from being reasonably performed in the human mind. Examiner respectfully disagrees with Applicant’s arguments. Examiner points to MPEP 2106.04(a)(2)(III)(C) which discloses that performing a mental process on a generic computer or computer environment or merely using a computer as a tool to perform a mental process still constitutes the mental process itself. That is, Examiner argues that a generic computer could be programmed to determine geometric coordinates of certain documents or text chunks in document(s), such as by electronic scanning or extraction of said document(s). Furthermore, assuming arguendo that said limitations were determined to fall outside the realm of being reasonably performed in the human mind, said limitations would constitute additional limitations and would be further considered under step 2A and 2B of the Alice/Mayo framework. That is, every limitation of this nature was determined to constitute insignificant, extra-solution activity and/or well-understood, routine, and/or conventional activity in prior art systems. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 19 of Arguments/Remarks that the Office Action’s characterization of the limitations reciting a mental process do not survive the amendments. More specifically, Applicant argues that generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node, wherein the generating comprises embedding, within the at least a portion of the extracted text, at least a portion of the structural metadata identifiers corresponding to the at least a portion of the extracted text” falls outside the realm of the human mind, because a human cannot embed coordinate-mapped structural metadata identifiers within text to form a machine-processable prompt payload. Examiner respectfully disagrees with Applicant’s arguments. Examiner points to MPEP 2106.04(a)(2)(III)(C) which discloses that performing a mental process on a generic computer or computer environment or merely using a computer as a tool to perform a mental process still constitutes the mental process itself. That is, Examiner argues that a generic computer could be programmed to determine geometric coordinates of certain documents or text chunks in document(s), such as by electronic scanning or extraction of said document(s), e.g. OCR which is well-known in prior art systems. And furthermore, generating an output of extracted content from a document based on information or other aspects of the content found in the document. And while Examiner generally agrees that determining coordinate-mapped structural metadata identifiers within text may not be performable by a human, this is typical of generic text and content extractors and in tandem with aspects of generating prompts based on said content, which IS performable by a human, this effectively represents a mental process under broadest reasonable interpretation. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 19-20 of Arguments/Remarks that the Office Action’s characterization of the limitations reciting a mental process do not survive the amendments. More specifically, Applicant argues that “providing the prompt payload to and receiving a structured response from the generative AI model” falls outside the realm of the human mind, because a human mind is not equipped to perform generative-AI inference or to produce a structured response that references coordinate-mapped structural metadata identifiers. Examiner respectfully disagrees with Applicant’s arguments. Examiner points to MPEP 2106.04(a)(2)(III)(C) which discloses that performing a mental process on a generic computer or computer environment or merely using a computer as a tool to perform a mental process still constitutes the mental process itself. That is, Examiner argues that a generic computer could be programmed to determine geometric coordinates of certain documents or text chunks in document(s), such as by electronic scanning or extraction of said document(s). While these aspects are then fed to a generative AI model, these limitations read as mere aspects of “apply it”. That is, a generic generative AI model is fed the prompt that could be generated by a human or generic computer, a black box occurs, and following this, content is outputted. That is, this is not specified to be a specialized generative AI model, and reads as applying off-the-shelf generative AI models but simply applied to patient care plans, etc. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 20 of Arguments/Remarks that the Office Action’s characterization of the limitations reciting a mental process do not survive the amendments. More specifically, Applicant argues that “determining whether the compatibility indicator satisfies the node capability specification” cannot be performed in the human mind, because at least one compatibility indicator is determined based on the capability specific of the service providing node. Examiner respectfully disagrees with Applicant’s arguments. Examiner points to MPEP 2106.04(a)(2)(III)(C) which discloses that performing a mental process on a generic computer or computer environment or merely using a computer as a tool to perform a mental process still constitutes the mental process itself. That is, a generic computer can be programmed to determining whether a compatibility indicator satisfies the node capability specification, such that the one compatibility indicator is determined based on the capability specific of the service providing node. As discussed above, a generic generative AI model is fed the prompt that could be generated by a human or generic computer, a black box occurs, and following this, content is outputted, and the compatibility indicators being generated by said model still reads as efforts of “apply it”, i.e. reads as applying off-the-shelf generative AI models but simply applied to patient care plans, etc. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 20 of Arguments/Remarks that the Office Action’s characterization of the limitations reciting a mental process do not survive the amendments. More specifically, Applicant argues that “generating the verification interface” requires resolving the specific structural metadata identifier referenced by the source citation against the geometric coordinates and rendering a visual overlay at the corresponding positions on the document image” which cannot be performed in the human mind. Examiner respectfully disagrees with Applicant’s arguments. Examiner points to MPEP 2106.04(a)(2)(III)(C) which discloses that performing a mental process on a generic computer or computer environment or merely using a computer as a tool to perform a mental process still constitutes the mental process itself. Examiner argues that a generic computer could be programmed to determine geometric coordinates of certain documents or text chunks in document(s), such as by electronic scanning or extraction of said document(s). And furthermore, generating an output of extracted content from a document based on information or other aspects of the content found in the document. And while Examiner generally agrees that determining coordinate-mapped structural metadata identifiers within text may not be performable by a human, this is typical of generic text and content extractors and in tandem with aspects of generating prompts based on said content, which IS performable by a human, this effectively represents a mental process, under broadest reasonable interpretation. While generating a visual interface may not be performable in the human mind, a generic computer can perform outputting a generic interface of results outputted from the generic, generative AI model. Furthermore, these aspects of outputting an interface amounts to insignificant, extra-solution activity and/or WURC of merely displaying gathered/analyzed information. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 20-21 of Arguments/Remarks that the system executes an OCR engine or a document layout analysis component and constructs a coordinate map that functions as a lookup table linking the extracted text to its physical position on the document image and therefore the limitations cannot be reasonably performed in the human mind. Examiner respectfully disagrees with Applicant’s arguments. Examiner points to MPEP 2106.04(a)(2)(III)(C) which discloses that performing a mental process on a generic computer or computer environment or merely using a computer as a tool to perform a mental process still constitutes the mental process itself. Examiner argues that a generic computer could be programmed to determine geometric coordinates of certain documents or text chunks in document(s), such as by electronic scanning or extraction of said document(s), such as performing OCR capabilities.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 21 of Arguments/Remarks that Examiner acknowledges that some of the limitations cannot be reasonably performed in the human mind, and therefore because they cannot be practically performed in the human mind, amended claim 1 does not recite a mental process. Examiner respectfully disagrees with Applicant’s arguments. Examiner argues that a generic computer could be programmed to determine geometric coordinates of certain documents or text chunks in document(s), such as by electronic scanning or extraction of said document(s). Furthermore, assuming arguendo that said limitations were determined to fall outside the realm of being reasonably performed in the human mind, said limitations would constitute additional limitations and would be further considered under step 2A and 2B of the Alice/Mayo framework. That is, every limitation of this nature was determined to constitute insignificant, extra-solution activity and/or well-understood, routine, and/or conventional activity in prior art systems. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 22-23 of Arguments/Remarks that claim 1 recites a specific improvement at least by providing a specific manner of displaying a limited set of information. Examiner respectfully disagrees with Applicant’s arguments. That is, this is not a specific or particular or improved interface, because prior art systems are able to perform and generally render said aspects found in the claims as well-understood, routine, and conventional activity. For instance, outputting results of analysis has been generally found to constitute insignificant, extra-solution activity, such as “Gathering and analyzing information using conventional techniques and displaying the result” as set forth by decisions in TLI Communications, 823 F.3d at 612-13, 118 USPQ2d at 1747-48. Furthermore, Core Wireless specifically set forth shortcomings of systems regarding certain aspects of the user interface that was then solved by the implementations found in Core Wireless. That is, Applicant does not set forth a technical problem regarding a user interface of the systems found in the prior art, but rather a lack of gathering specific types of information to be communicated and/or presented using said interfaces, which is a wholly different problem that is not related to technological components themselves, but is instead related to abstract notions of gathering and presenting results of content extracted from documents. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 23 of Arguments/Remarks that there is a direct nexus between the claimed technical features and the interface improvement, because the interface improvement is enabled by the claimed identifier-embedding-and-resolution mechanism found in amended claim 1, i.e. the structural metadata identifiers are “embedded within the at least a portion of the extracted text” of the prompt payload. Examiner respectfully disagrees with Applicant’s arguments. Merely providing a specified version of an interface, such as electing which content is presented in an interface based on applying a generic, generative AI model without further describing the algorithm for determining which content to be selected or how the selection is an improvement in developing said interfaces beyond the generative AI representing a substantially black box does not amount to an interface improvement. That is, Applicant does not set forth a technical problem regarding a user interface of the systems found in the prior art, but rather a lack of gathering specific types of information to be communicated and/or presented using said interfaces, which is a wholly different problem that is not related to technological components themselves, but is instead related to abstract notions of gathering and presenting results of content extracted from documents. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 23-24 of Arguments/Remarks that the Specification confirms that the claimed mechanism yields concrete technological improvements, such as identifying the technical effects of the claimed positional-correction and visual-overlay system including “enhancing interpretability in healthcare decision workflows”, “future validation and compliance verification”, “enabling explainable and verifiable AI decision-making”, “ensuring consistent and traceable mapping between unstructured text and structured AI output”, etc., and therefore constitutes a practical application under the Alice/Mayo framework. Examiner respectfully disagrees with Applicant’s arguments. Examiner argues that these are not “concrete” technological improvements. That is, none of these improvements improve upon concrete or physical components implementing the abstract idea at hand. Rather, each of these aspects improve upon abstract concepts, such as enhancing interpretability, validation and verification which are all human/abstract concepts/ideas. And therefore, improvements to these aspects represent improvements to the already-characterized abstraction, rather than improvements to the technology itself. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 24-25 of Arguments/Remarks that Examiner’s positions reflect two legal errors: 1) that eligibility does not require the specification to recite particular technical effects of the claimed technology and 2) the claim is not a mere “tailoring” of outputs, as characterized by the Examiner. Examiner respectfully disagrees with Applicant’s arguments. Regarding 1), while Applicant is generally correct in stating that eligibility does not require the specification to recite particular technical effects of the claimed technology, Examiner argues that while it is not required to recite particular technical effects in the Specification, if this aspect is not explicitly set forth, then Applicant is merely relying on knowledge of the Examiner or general knowledge in the state of the art at the time of filing to convey said technical effects unless they are otherwise specified. Therefore, while Applicant may have particular technical effects that are intended to be conveyed, without further detailing or specifying said aspects, said effects may not be as apparent. Regarding 2) while Examiner concedes that a model is fed a prompt in order to analyze and produce an output based on said analysis, these efforts read as mere efforts to apply a generic, well-known generative AI model for purposes of creating indications relating to scanned or extracted text/content. That is, these aspects relate substantially to receiving data, analyzing data, and outputting results/indications of said results, but recited for a generative AI or automated fashion. While Applicant argues in view of McRO v. Bandai Namco Games Am. Inc., Applicant does not give specific arguments or substantiation regarding this precedential case. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 25-26 of Arguments/Remarks that the additional elements must be considered as an ordered combination, and that this combination is not a generic or conventional arrangement and it is the specific arrangement that enables the coordinate-precise, source-attributed verification interface. Applicant further argues that the non-conventional, non-generic arrangement supplies an inventive concept under Step 2B in view of BASCOM. Examiner respectfully disagrees with Applicant’s arguments. Each of the additional elements, alone or in combination, were determined to represent well-understood, routine, and/or conventional activity found in prior art systems. That is receiving data/prompts, performing analysis using a generic, off-the-shelf generative AI model, performing OCR/extracted text efforts, and outputting said efforts with annotations therein via an interface all represent well-understood, routine, and/or conventional activity found in prior art systems, but simply applied for care plans or medical data of a patient. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 27 of Arguments/Remarks that the Office Action provides no evidence that the claimed combination is well-understood, routine, or conventional. Examiner respectfully disagrees with Applicant’s arguments. In the “Claim Rejections – 35 U.S.C. 101” section of the current and previous Office Actions, each additional element, individually and in combination, was determined to represent well-understood, routine, and conventional activity found in prior art systems, such as via Berkheimer evidence. Following the portions representative of the Berkheimer evidence pointing towards the elements representing well-understood, routine, and conventional activity found in prior art systems, the Office Action clearly states “Looking at the limitations as an ordered combination adds nothing that is not already present when looking at the elements taken individually. There is no indication that the combination of elements improves the functioning of a computer or improves any other technology. Their collective functions merely provide conventional computer implementation”. Therefore, evidence of the limitations representing ell-understood, routine, and conventional activity found in prior art systems is provided, and therefore the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 27-28 of Arguments/Remarks that Examiner’s interview comment that the comment does not recite a new or improved AI model does not bear on the Step 2B analysis, such that “the search for an inventive concept should not be confused with a novelty or non-obviousness determination”. And therefore, the ordered combination represents significantly more. Examiner respectfully disagrees with Applicant’s arguments. While Examiner generally agrees that prior art rejections, i.e. novelty/obviousness, should not be conflated with the determinations of an inventive concept, Examiner was referencing prior art systems providing evidence to the additional elements constituting well-understood, routine, and/or conventional activity in prior art systems. Furthermore, assuming arguendo that Examiner was referencing said aspects of novelty or non-obviousness determinations, comments made in an interview are not substantive unless otherwise noted or directly addressed in an Office Action and rejections are changed as a result. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 101 rejections of claims 1-20, Applicant argues on p. 28-29 of Arguments/Remarks that because independent claims 1, 6, & 25 purportedly represent patent-eligible subject matter, that dependent claims, by virtue of dependency, should also represent patent eligible-subject matter. Examiner respectfully disagrees with Applicant’s arguments. As discussed above, the independent claims were determined to still represent patent-ineligible subject matter, and therefore Applicant’s arguments are rendered moot, because the independent claims do not represent patent-eligible subject matter. Therefore the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 29-31 of Arguments/Remarks that previously cited portions of Couleaud and Wang do not read on independent claims 1, 6, & 25. Examiner agrees with Applicant’s arguments. Therefore, a new ground of rejection has been made under 35 U.S.C. 103 over Couleaud and Wang. Newly reasoned/cited portions of Couleaud and Wang are applied to read on the newly amended limitations found in independent claims 1, 6, & 25. As such, independent claims and claims dependent therefrom remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 31-32 of Arguments/Remarks that amended claim 1 extracts text from a received document image and renders evaluation results onto that image, while Couleaud generates new images from a user-supplied text prompt. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. Examiner contends that it is the combination of Couleaud and Wang that reads on the entirety of amended claim 1. That is, Couleaud Par [0078] & [0090] discloses an image capture system pre-processing image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image; Couleaud Par [0044] discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values. Therefore, Couleaud effectively performs extraction of text and generation of an interface of said AI evaluation regarding various media found in the extraction, however Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media items. Furthermore, Wang Par [0055] which discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051], and Wang Par [0131] discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output. Therefore, it is the combination of Couleaud and Wang which discloses the entirety of claim 1. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 32-33 of Arguments/Remarks that Couleaud does not disclose embedding structural metadata identifiers within the extracted text of the prompt payload. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. Couleaud Par [0044] discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata identifier, and/or encoding information regarding a depth of an object, which would include text extraction therein and/or metadata identifiers, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values. Therefore, while not necessarily “embedded”, it is understood that said vector that specifies pixel value information, i.e. metadata identifier, and/or encoding information regarding a depth of an object, would include text extraction therein and/or metadata identifiers, and therefore constitute embedding said aspects within the prompt provided to the generative AI for content extraction. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 33 of Arguments/Remarks that Couleaud does not disclose a structured response in which a source citation references a specific structural metadata identifier. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. Couleaud Par [0044] discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata identifier, and/or encoding information regarding a depth of an object, which would include text extraction therein and/or metadata identifiers, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values. Therefore, it is understood that said vector that specifies pixel value information, i.e. metadata identifier, and/or encoding information regarding a depth of an object, would include text extraction therein and/or metadata identifiers, and therefore constitute embedding said aspects within the prompt provided to the generative AI for content extraction. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 33-34 of Arguments/Remarks that Couleaud does not disclose a verification interface that renders a visual overlay by resolving the identifier referenced by the source citation against the geometric coordinates. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. The combination of Couleaud and Wang effectively discloses a verification interface that renders a visual overlay by resolving the identifier referenced by the source citation against the geometric coordinates. That is, the verification portion regarding Couleaud Par [0044] discloses various object extraction from an image, including a vector that specifies pixel value information, i.e. metadata, and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090], which would include geometric coordinates of the received image, and further discloses generating a map associating image and object with associated indicators of pixel values. Therefore, Couleaud effectively performs extraction of text and generation of an interface of said AI evaluation regarding various media found in the extraction, however Wang Par [0049]-[0051] which discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media items. Furthermore, Wang Par [0055] discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051], and Wang Par [0131] discloses retrieval augmented generation (RAG) component for retrieving additional information to be used as part of the input or prompt, such that an input processor communicating with in order to identify relevant text and/or other data to provide to the generative LM as additional context or sources of information from which to identify the response, answer, or output. Therefore, the combination of Couleaud and Wang effectively discloses a verification interface that renders a visual overlay by resolving the identifier referenced by the source citation against the geometric coordinates. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 34 of Arguments/Remarks that Couleaud does not disclose “generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node”. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. Couleaud Par [0050] & Fig. 1 discloses generating a set of textual prompts, i.e. a structured response, corresponding to a received textual prompt and image associated with said textual prompt. Couleaud Par [0077] discloses “node capability specifications” at least by each layer may comprise one or more nodes that may be associated with learned parameters (e.g., weights and/or biases), and/or connections between nodes may represent parameters (e.g., weights and/or biases) learned during training (e.g., using backpropagation techniques, and/or any other suitable techniques), and in some embodiments, the nature of the connections may enable or inhibit certain nodes of the network, constituting node capabilities. Couleaud Par [0062] discloses creation of composite images, each being a variation of a multi-layer image generated based on text, that uses various image position, size, color, texture, appearance, etc., i.e. thereby utilizing location data of the document image. Couleaud Par [0078] & [0090] discloses pre-processing to be performed, such that an image capture system may pre-process image or text data to be input into the trained machine learning model, such that an image-to-text model may be used to generate a second plurality of objects, i.e. text objects since the model is image-to-text, from each image. Couleaud Par [0044] discloses various object extraction from an image, including a vector that specifies pixel value information and/or encoding information regarding a depth of an object, which would include text extraction therein, as specified by Couleaud Par [0078] & [0090]. Finally, Couleaud Par [0002] which specifically mentions “embedding” given text caption. Therefore, it is understood that Couleaud effectively discloses generating a prompt payload comprising at least a portion of the extracted text and a node capability specification of a service providing node. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 35-36 of Arguments/Remarks that Couleaud does not disclose “at least one compatibility indicator determined based on the capability specification of the service providing node”. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. The combination of Couleaud and Wang effectively discloses “at least one compatibility indicator determined based on the capability specification of the service providing node”, not just Couleaud. That is, Couleaud Par [0074] discloses the image processing system may input data into a model to generate outputs, such that said outputs can be adjusted based on ground truth values, e.g. annotated indications of the correct or desired outputs for given input(s), and further discloses the training process may be repeated until results stop improving or a certain performance level is achieved, i.e. understood to constitute a “compatibility indicator”. However, Wang Par [0049]-[0051] discloses generating media features via a media model, such that any text model and media model may include one or more self-attention blocks to identify associations, i.e. compatibility, between different units of the respective inputs, i.e. text inputs and media inputs, such that text features and media features may be processed by a multi-modal transformer that uses one or more cross-attention blocks (but may also include any number of self-attention blocks) to identify associations between units of prompt and units of content of media item. Wang Par [0055] discloses generating probabilities that a particular unit of a media item is associated with an individual object based on said text and media features established in Wang Par [0049]-[0051] and therefore bolsters aspects of the metrics produced in Couleaud being a “compatibility indicator” in particular. Therefore, the combination of Couleaud and Wang effectively discloses a “at least one compatibility indicator determined based on the capability specification of the service providing node” under broadest reasonable interpretation. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 36-38 of Arguments/Remarks that none of the secondary references cure the deficiencies of Couleaud. Therefore, the 35 U.S.C. 103 rejections should be withdrawn. Examiner respectfully disagrees with Applicant’s arguments. The secondary references do not have to cure the argued deficiencies, because Couleaud effectively discloses said limitations argued by Applicant. As such, the claims remain rejected under 35 U.S.C. 103.
Regarding 35 U.S.C. 103 rejections of claims 1-20, Applicant argues on p. 36-38 of Arguments/Remarks that because independent claims 1, 6, & 25 are purportedly allowable over the prior art, that by virtue of dependency, claims dependent from claims 1, 6, & 25 are also allowable over the prior art. Examiner respectfully disagrees with Applicant’s arguments. As discussed above, the independent claims were determined to still represent patent-ineligible subject matter, and therefore Applicant’s arguments are rendered moot, because the independent claims do not represent patent-eligible subject matter. Therefore, the claims remain rejected under 35 U.S.C. 101, because the claims effectively recite a mental process under broadest reasonable interpretation, and do not amount to a practical application or significantly more than the recited abstract idea.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Vasylyev et al. (U.S. Patent Publication No. 2024/0412720) discloses providing a contextualized response to a user using AI, such that the conversational context data is updated with a tokenized representation of a generated response, and output the response;
Cook et al. (U.S. Patent Publication No. 2024/0126794) discloses receiving at least one user query from a user and generating a query response as a function of the at least one user query and the contextual data using the digital assistant, and displaying said query response using the digital assistant on a device/interface.
Applicant's amendment necessitated the new grounds 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 HUNTER J RASNIC whose telephone number is 571-270-5801. The examiner can normally be reached M-F 8am-5:30pm.
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, Shahid Merchant can be reached at (571) 270-1360. 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.
/H.R./Examiner, Art Unit 3684
/Shahid Merchant/Supervisory Patent Examiner, Art Unit 3684
Prosecution Insights