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 .
This office action is in response to the Request for Continued Examination filed on 18 March 2026.
This office action is made Non Final.
Claims 1, 8, and 15 have been amended.
Claims 7 and 14 have been cancelled.
The objection to the abstract, claim objections and art rejections from the previous office action has been withdrawn as neccessited.
Claims 1,4, 8, 11, 15, 18 are pending. Claims 1, 8, and 15 are independent claims.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 3/18/26 has been entered.
Specification
After reevaluation of the amended Abstract filed on 10/30/2025, the amended Abstract has been accepted and the objection to the abstract has been withdrawn.
Claim Interpretation
The limitations of Claim 15 no longer invoke 35 U.S.C. § 112(f) (pre-AIA 35 U.S.C. 112, sixth paragraph) in response to Applicant’s amendments.
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.
Claim(s) 1, 4, 8, 11, 15, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Wilson et al (US20230052372, EFD 8/13/2021) in further view of Atmakuri et al (US 20020188585) in further view of Bhandarkar et al (US20220035993, EFD 2020) in further view of Tanikawa (US20160063064, 2016) in further view of Elliman (US20200134083, 2020) in further view of Matiukhov (US 20220156490, 5/2022)
As per independent claim 1, Wilson et al discloses a method comprising:
obtaining a data extraction request associated with a document; (0074, 0082, 0128: uploading document for data to be extracted; form of a request)
in response to obtaining the request: generating a data extraction prediction using a prediction model and the document; (0007; Claim 10: suggested texts, associated with a label, are selected)
providing the data extraction prediction to a user via a graphical user interface (GUI); (FIG 14; 0087)
generating, by a user, a user validation using interaction capabilities of the GUI (FIG 14; 0087; 0008: user interacts with the GUI)
wherein: the user validation includes indicators that specify portions of the data extraction prediction that are correct or incorrect (0056; 0074: user accepts or corrects prediction/suggested text)
the interaction capabilities of the GUI comprise of: clicking a button; (0084, 0087; FIG 10, 14)
entering in information via a keyboard to provide feedback on portions of the data extraction prediction (0008: second user input to correct the suggested label so that it correctly describes the selection of text, see also 0143. 0135 discloses keyboard can be a device that profiles input. One of a skilled artisan would have realized that providing a correct description of the select text would involve the use of a keyboard to enter information (i.e. description) that provides feedback for an incorrect prediction which is corrected)
obtaining the user validation associated with the data extraction prediction from the user; making a determination that the user validation indicates that the data extraction prediction is not correct based on indicators; (0056, 0087, 0108,0143; accept or reject/correct prediction)
in response to the determination: generating an updated data extraction prediction based on the user validation; (0007,0056, 0108: User provides a correction, a form of generating an updated prediction)
performing data preparation on a copy of the document based on the user validation and the updated data extraction prediction to generate a training document (0073-0074: User corrects the suggest text/label by annotating the document (copy of the document) to create training data used to train the models (0010: the stored annotated electronic documents as training data to train a machine learning model to automatically predict and extract selections of text corresponding to one or more labels from non-annotated electronic documents ))
storing the training document, the user validation, and the data extraction prediction in a document repository, wherein the document repository comprises: a plurality of previously generated training documents, a plurality of previously generated data extraction predictions, and a plurality of previously obtained user validations. (0007-0008, 0074: The annotated data includes accepted and corrected predicted annotations made by the user wherein the predictions of text corresponding to the one or more labels are graphically highlighted. The annotated data is then stored with the document itself. (0008: storing the suggested label, the selection of text, and the location for the selection of text within the electronic document as an annotated electronic document if the suggested label correctly describes the selection of text) Thus, the annotated document includes the original document, the generated predictions, and the feedback (accepting/correcting) of the predictions. Furthermore, one of a skilled artisan would have realized since the annotated document was being stored, then a form of a memory is needed to provide in order to store the annotated documents. Therefore, Wilson discloses a form of a document repository. In addition, Wilson et al in 0012 discloses repeating the method of annotating one or more additional (non-annotated) electronic documents and storing the one or more additional annotated electronic documents. Thus, Willson indicates a plurality of annotated electronic documents, as presented by Wilson, are stored in the form of a memory.)
initiating performance of additional document processing using the document based on the updated data extraction prediction. (0073-0074, 0109, 0111, 0128: The corrected predictions are used to retrain the prediction model on the document and then run the data extraction model; A cycle of annotating additional electronic documents to generate additional training data, training the one or more data extraction models, and validating them against additional sets of validation documents, may be repeated until a desired level of data extraction accuracy is achieved.)
monitoring, by a document data extractor, points in time for a prediction model update for a model prediction event, (0097: user decides that an existing model should be update; a form of point in time)
monitoring further comprises continuously calculating an accuracy of the prediction model (0074: discloses training the one or more data extraction models, and validating them against additional sets of validation documents, may be repeated until a desired level of data extraction accuracy is achieved. Once the desired performance metrics for the trained extraction model(s) have been achieved, the one or more trained extraction models may be deployed; 0096: The performance of the term-based data extraction model may then be compared to a set of target performance metrics (e.g., annotation accuracy, etc.) at step 2218 to decide whether the model should be updated (e.g., further trained) or deployed for use. Thus, 0074, 0096 discloses repeating checking metrics/accuracy of the model to determine if the model is ready for use or be retrained.
identifying, based on the monitoring/determination, a prediction model update event; (FIG 22; 0097: user decides that an existing model should be updated at step 2218 that results in steps 2210, 2212, 2214, and 2216 repeating which includes display of the annotation results for user review and feedback (which is explain in 0007-0008, 0074)
obtaining training documents from the document repository; (0008: using the stored annotated electronic documents as training data to train a machine learning model. As explained above, one of a skilled artisan would have realized since the annotated document was being stored, then a form of a memory is needed to provide in order to store the annotated documents. Therefore, Wilson discloses a form of a document repository. Thus, using the stored annotated documents would resulted in obtaining the documents from the repository)
performing data preparation on the training documents to generate updated training documents, wherein the updated training documents comprise: the plurality of previously generated training documents, the plurality of previously generated data extraction predictions, and the plurality of previously obtained user validations; (0007-0008, 0074: cycle of annotating additional electronic documents to generate additional training data. 0097 discloses an expanded set of labeled training data is compiled at step 2210 (e.g., using the annotated documents and associated label and annotation data generated at step 2206, and/or using additional labeled training data), and the automated data extraction model is trained (or re-trained) at step 2212.( form of data preparation on the documents) The trained or re-trained model is then used to process remaining or additional training / validation documents at step 2214, followed by display of the annotation results for user review and feedback at step 2216. Step 2216 involves repeating the steps that occur in 0007-0008, 0074,0087 which includes the user correcting the suggested label so that it correctly describes the selection of text. As explained in Claim 5, each annotated document includes the previous version of that document, the generated predictions, and the feedback (accepting/correcting) of the predictions. Thus, these stored annotated documents are obtained and used to re-train the model.)
generating an updated prediction model using the updated training documents; and replacing the prediction model with the updated prediction model. (0074, 0097, 0128: model is re-trained/tuned, a form of updating the model (replacing the old model with a new model))
While Wilson discloses the client received the documents via uploading (0074, 0082, 0128) (a form of obtaining a data extraction request associated with the document), the cited art fails to specifically disclose wherein the extraction request includes contact information associated with a third-party entity that holds the document, wherein the contact information includes :a network address; an internet protocol (IP) address ;a host address; digital signatures; and a public key infrastructure information. In other words, the cited art fails to disclose a request, for a document (for data extraction), comprising contact information associated with a third-party entity wherein the contact information includes :a network address; an internet protocol (IP) address; a host address; digital signatures; and a public key infrastructure information.
However, based on the 112b rejection, Atmakuri et al discloses a requesting a server for a document wherein the request includes the network address for a server. In response to the request, the server obtains the requested document from a database and returns it to the client. (0021, 0027) Thus, Atmakuri et al discloses/analyze a network address for the server’s location. The Examiner states the additional elements listed in the claims: i.e. an internet protocol (IP) address; a host address; digital signatures; a public key; and an infrastructure information are nominally recited and the claims do not distinguish how each of these elements are used for contacting the third-party entity.
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the disclosed feature(s) of Atmakuri et al’s ability to send a document request to a server using a particular type of contact information since it would have provided the intrinsic advantage of establishing a secure, verifiable, and legally admissible audit trail. This address acts as a digital "return address" for data routing and provides evidence of where and when a document was accessed or signed.
Furthermore, Atmakuri et al does not expressly show the contact information also includes an internet protocol (IP) address ;a host address; digital signatures; a public key; and an infrastructure information.
However, these differences are only found in the nonfunctional descriptive material and are not functionally involved in the steps recited. The obtaining steps would be performed the same regardless of the data. Thus, this descriptive material will not distinguish the claimed invention from the prior art in terms of patentability, see In re Gulack, 703 F.2d 1381, 1385, 217 USPQ 401, 404 (Fed. Cir. 1983); In re Lowry, 32 F.3d 1579, 32 USPQ2d 1031 (Fed. Cir. 1994).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of Applicant’s invention to have the request include any kind of contact information because such data does not functionally relate to the steps in the method claimed and because the subjective interpretation of the data does not patentably distinguish the claimed invention.
Therefore, the combination of Wilson et al and Atmakuri et al teaches the subject matter of obtaining a data extraction request associated with a document; wherein the extraction request includes contact information associated with a third-party entity that holds the document, wherein the contact information includes :a network address; an internet protocol (IP) address ;a host address; digital signatures; and a public key infrastructure information.
Furthermore, the cited art fails to specifically disclose generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the user validation includes UI indicators that specify portions of the data extraction prediction that are correct and incorrect, highlighting portions of the data extraction prediction with a mouse; and making a determination that the user validation indicates that the data extraction prediction is not correct based on the UI indicators. However, Bhandarkar et al discloses supervised or semi-supervised machine learning to generate the training dataset necessary to train the machine learned model. (0036) The model identifies a set of passages in the document to highlight wherein the highlighted text passage is designated by a visual indicator. (0035) Furthermore, the user is able to provide feedback regarding the correctness of each highlighted text passage. (0019, 022: the user may approve or reject one or more highlighted text passages identified and automatically highlighted) The user may delete one or more of the highlighted text passages 280 (for instance, in embodiments where the automatic highlight is incorrect). (0041) In other words, the user is able to indicate if a highlight text passage is correct or incorrect by maintaining the text passage being highlight or removing highlight from the highlight text passage. Thus, the presence of a text remains being highlighted indicates the portions of data are correct wherein if the text is unhighlighted indicates the portions are incorrect. This is a form of user validation. Furthermore, additional text passages can be highlighted. (0031, 0041) Bhandarkar discloses the user is able to manually highlight additional passages by selecting the text (user selection) (0031: The user may highlight and/or select candidate text passages for highlighting by receiving user input). While Bhandarkar et al does not explicitly state that a mouse was use to perform the manual highlighting of text by the user, a selection using an input device such as a mouse is well known in the art. The Examiner takes OFFICIAL NOTICE on this fact that it is typically known a mouse is used for selecting and highlighting text; therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified Bhandarkar’s ability to allow a user to provide an input to make a highlighting selection, such that the input selection could have been through an input device of a mouse as well known in the art. This combination would have provided the intrinsic advantage excels at selecting arbitrary or non-contiguous blocks of text with high precision. In addition, based on the unhighlighting or new highlighting by the user, the machine learning model is re-trained. (0041) Thus, Bhandarkar is able to determine if the user validation indicates identified text passage by the model is correct or incorrect if the text passage remains highlight (model incorrect identified the text passage when its highlight indicator is removed) Thus, Bhandarkar discloses in response to determining that a highlighted text passages is incorrect by its UI indicator (no longer highlight), the machine learning model for identifying relevant text (by highlighting) is re-trained.
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Bhandarkar since it would have provided the benefit of an improved, efficient, and more reliable document execution experience for the user.
Thus, in conjunction with Wilson et al, the combination teaches generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the user validation includes UI indicators that specify portions of the data extraction prediction that are correct and incorrect, the interaction capabilities of the GUI comprise of: clicking a button; entering in information via a keyboard to provide feedback on portions of the data extraction prediction; and highlighting portions of the data extraction prediction with a mouse;
In addition, the cited art fails to specifically disclose generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the interaction capabilities of the GUI comprise of: checking a box. However, Tanikawa discloses generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the interaction capabilities of the GUI comprise of: checking a box (0028: user checks a box to indicate data has been validated)
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Tanikawa since checking a box would have provided the intrinsic advantage of provides a clear, simple, visual way to confirm an item meets a specific condition or status, making data more reliable and easier to analyze.
However, the cited art fails to specifically discloses monitoring, by a document data extractor, points in time on a prediction model update schedule for a model prediction update event, wherein the document data extractor is a physical hardware processor. However, Ellison discloses the data classification system 100 (e.g., classification computing entity 65) continuously retrains the machine learning models and applies the retrained models to new data assets, wherein the data classification system 100 can retrain the one or more machine learning models on a fixed schedule (e.g., hourly, daily, weekly, and/or the like). (0076-0077) 0028 discloses the classification computing entity 65 includes a processor. Thus, Ellison discloses a physical hardware processor that retrains the model on a fixed schedule, a form of monitoring by a “document data extractor”)
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Ellison since it would have provided the intrinsic advantage of improve accuracy over time and ensures the model periodically "re-learns" the current patterns.
Furthermore, as explained above, Wilson et al disclose continuously calculating an accuracy of the prediction model (0074: discloses training the one or more data extraction models, and validating them against additional sets of validation documents, may be repeated until a desired level of data extraction accuracy is achieved. Once the desired performance metrics for the trained extraction model(s) have been achieved, the one or more trained extraction models may be deployed; 0096: The performance of the term-based data extraction model may then be compared to a set of target performance metrics (e.g., annotation accuracy, etc.) at step 2218 to decide whether the model should be updated (e.g., further trained) or deployed for use. In other words, It appears that a form of a threshold was used since the model gets retrained until a desired level of data extraction accuracy is achieved. However, Wilson fails to specifically disclose making, based on the monitoring, a second determination that the accuracy is below a threshold identifying, based on the second determination, a prediction model update event. However, Matiukhov discloses a determination is made as to whether the predictions made by the model are below an accuracy threshold. In addition, Matiukhov discloses that if the model is below the explicit threshold, then the model is retrained and this process is repeated until it reaches the desired threshold. Thus, Matiukhov also discloses (0064-0065; Claim 8)
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Matiukhov since it would have provided the benefit an automated document data processing that is time-saving less prone to error. (0022)
As per dependent claim 4, Wilson et al discloses performing the data preparation on the copy of the document based on the user validation and the updated data extraction prediction to generate the training document: generating a spatial feature associated with the updated data extraction prediction; and generating a textual feature associated with the updated data extraction prediction. (0055-0056; 0073-0074: User corrects the suggest text/label by annotating the document to create training data used to train the models. The correction results in location of the selection and the text of selected (generated) stored)
As per independent claim 8, Claim 8 recites similar limitations as in Claim 1 and is rejected under similar rationale. Furthermore, Wilson et al discloses a medium (0138; FIG 30)
As per dependent claim 11, Claim 11 recites similar limitation(s) as in Claim 4 and is rejected under similar rationale.
As per independent claim 15, Claim 15 recites similar limitations as in Claim 1 and is rejected under similar rationale. Furthermore, Wilson et al discloses
a plurality of clients (0020, 0068, 0073; FIG 1: discloses a plurality of clients/users performing the functionality of Wilson. Thus, each user is a single client of the plurality of clients/users)
In addition, Wilson discloses the following functionalities:
document processing engine (0137: software that includes the programming that embodies the functionality of Wilson; FIG 1 ) executed by the processor on a computing device (0134, 0137: executed by the processor of computing device)
obtaining a data extraction request associated with a document from a client of the plurality of clients; (0074, 0128)
in response to obtaining the request: generate a data extraction prediction using a prediction model and the document; (0007; Claim 10: suggested texts, associated with a label, are selected)
provide the data extraction prediction to a user via a graphical user interface (GUI); (FIG 14; 0087)
generate, by a user, a user validation using interaction capabilities of the GUI (FIG 14; 0087; 0008: user interacts with the GUI)
wherein: the user validation includes indicators that specify portions of the data extraction prediction that are correct or incorrect (0056; 0074: user accepts or corrects prediction/suggested text)
the interaction capabilities of the GUI comprise of: clicking a button; (0084, 0087; FIG 10, 14)
entering in information via a keyboard to provide feedback on portions of the data extraction prediction (0008: second user input to correct the suggested label so that it correctly describes the selection of text, see also 0143. 0135 discloses keyboard can be a device that profiles input. One of a skilled artisan would have realized that providing a correct description of the select text would involve the use of a keyboard to enter information (i.e. description) that provides feedback for an incorrect prediction which is corrected)
obtain the user validation associated with the data extraction prediction from the user; make a determination that the user validation indicates that the data extraction prediction is not correct based on indicators; (0056, 0087, 0108,0143; accept or reject/correct prediction)
in response to the determination: generate an updated data extraction prediction based on the user validation; (0007,0056, 0108: User provides a correction, a form of generating an updated prediction)
perform data preparation on a copy of the document based on the user validation and the updated data extraction prediction to generate a training document (0073-0074: User corrects the suggest text/label by annotating the document (copy of the document) to create training data used to train the models (0010: the stored annotated electronic documents as training data to train a machine learning model to automatically predict and extract selections of text corresponding to one or more labels from non-annotated electronic documents ))
storing the training document, the user validation, and the data extraction prediction in a document repository, wherein the document repository comprises: a plurality of previously generated training documents, a plurality of previously generated data extraction predictions, and a plurality of previously obtained user validations. (0007-0008, 0074: The annotated data includes accepted and corrected predicted annotations made by the user wherein the predictions of text corresponding to the one or more labels are graphically highlighted. The annotated data is then stored with the document itself. (0008: storing the suggested label, the selection of text, and the location for the selection of text within the electronic document as an annotated electronic document if the suggested label correctly describes the selection of text) Thus, the annotated document includes the original document, the generated predictions, and the feedback (accepting/correcting) of the predictions. Furthermore, one of a skilled artisan would have realized since the annotated document was being stored, then a form of a memory is needed to provide in order to store the annotated documents. Therefore, Wilson discloses a form of a document repository. In addition, Wilson et al in 0012 discloses repeating the method of annotating one or more additional (non-annotated) electronic documents and storing the one or more additional annotated electronic documents. Thus, Willson indicates a plurality of annotated electronic documents, as presented by Wilson, are stored in the form of a memory.)
initiate performance of additional document processing using the document based on the updated data extraction prediction. (0073-0074, 0109, 0111, 0128: The corrected predictions are used to retrain the prediction model on the document and then run the data extraction model)
monitoring, by a document data extractor, points in time for a prediction model update for a model prediction event, (0097: user decides that an existing model should be update; a form of point in time)
monitoring further comprises continuously calculating an accuracy of the prediction model (0074: discloses training the one or more data extraction models, and validating them against additional sets of validation documents, may be repeated until a desired level of data extraction accuracy is achieved. Once the desired performance metrics for the trained extraction model(s) have been achieved, the one or more trained extraction models may be deployed; 0096: The performance of the term-based data extraction model may then be compared to a set of target performance metrics (e.g., annotation accuracy, etc.) at step 2218 to decide whether the model should be updated (e.g., further trained) or deployed for use. Thus, 0074, 0096 discloses repeating checking metrics/accuracy of the model to determine if the model is ready for use or be retrained. )
identifying, based on the monitoring/determination, a prediction model update event; (FIG 22; 0097: user decides that an existing model should be updated at step 2218 that results in steps 2210, 2212, 2214, and 2216 repeating which includes display of the annotation results for user review and feedback (which is explain in 0007-0008, 0074)
obtain training documents from the document repository; (0008: using the stored annotated electronic documents as training data to train a machine learning model. As explained above, one of a skilled artisan would have realized since the annotated document was being stored, then a form of a memory is needed to provide in order to store the annotated documents. Therefore, Wilson discloses a form of a document repository. Thus, using the stored annotated documents would resulted in obtaining the documents from the repository)
perform data preparation on the training documents to generate updated training documents, wherein the updated training documents comprise: the plurality of previously generated training documents, the plurality of previously generated data extraction predictions, and the plurality of previously obtained user validations; (0007-0008, 0074: cycle of annotating additional electronic documents to generate additional training data. 0097 discloses an expanded set of labeled training data is compiled at step 2210 (e.g., using the annotated documents and associated label and annotation data generated at step 2206, and/or using additional labeled training data), and the automated data extraction model is trained (or re-trained) at step 2212.( form of data preparation on the documents) The trained or re-trained model is then used to process remaining or additional training / validation documents at step 2214, followed by display of the annotation results for user review and feedback at step 2216. Step 2216 involves repeating the steps that occur in 0007-0008, 0074,0087 which includes the user correcting the suggested label so that it correctly describes the selection of text. As explained above, each annotated document includes the previous version of that document, the generated predictions, and the feedback (accepting/correcting) of the predictions. Thus, these stored annotated documents are obtained and used to re-train the model.)
generate an updated prediction model using the updated training documents; and replace the prediction model with the updated prediction model. (0074, 0097, 0128: model is re-trained/tuned, a form of updating the model (replacing the old model with a new model))
Since the cited art et al discloses the functionalities that are performed by the engine, then the cited art et al discloses the necessary document processing engine in order to perform the functionalities.
While Wilson discloses the client received the documents via uploading (0074, 0082, 0128) (a form of obtaining a data extraction request associated with the document), the cited art fails to specifically disclose wherein the extraction request includes contact information associated with a third-party entity that holds the document, wherein the contact information includes :a network address; an internet protocol (IP) address ;a host address; digital signatures; and a public key infrastructure information. In other words, the cited art fails to disclose a request, for a document (for data extraction), comprising contact information associated with a third-party entity wherein the contact information includes :a network address; an internet protocol (IP) address; a host address; digital signatures; and a public key infrastructure information.
However, based on the 112b rejection, Atmakuri et al discloses a requesting a server for a document wherein the request includes the network address for a server. In response to the request, the server obtains the requested document from a database and returns it to the client. (0021, 0027) Thus, Atmakuri et al discloses/analyze a network address for the server’s location. The Examiner states the additional elements listed in the claims: i.e. an internet protocol (IP) address; a host address; digital signatures; a public key; and an infrastructure information are nominally recited and the claims do not distinguish how each of these elements are used for contacting the third-party entity.
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the disclosed feature(s) of Atmakuri et al’s ability to send a document request to a server using a particular type of contact information since it would have provided the intrinsic advantage of establishing a secure, verifiable, and legally admissible audit trail. This address acts as a digital "return address" for data routing and provides evidence of where and when a document was accessed or signed.
Furthermore, Atmakuri et al does not expressly show the contact information also includes an internet protocol (IP) address ;a host address; digital signatures; a public key; and an infrastructure information.
However these differences are only found in the nonfunctional descriptive material and are not functionally involved in the steps recited. The obtaining steps would be performed the same regardless of the data. Thus, this descriptive material will not distinguish the claimed invention from the prior art in terms of patentability, see In re Gulack, 703 F.2d 1381, 1385, 217 USPQ 401, 404 (Fed. Cir. 1983); In re Lowry, 32 F.3d 1579, 32 USPQ2d 1031 (Fed. Cir. 1994).
Therefore, it would have been obvious to a person of ordinary skill in the art before the effective filing date of Applicant’s invention to have the request include any kind of contact information because such data does not functionally relate to the steps in the method claimed and because the subjective interpretation of the data does not patentably distinguish the claimed invention.
Therefore, the combination of Wilson et al and Atmakuri et al teaches the subject matter of obtaining a data extraction request associated with a document; wherein the extraction request includes contact information associated with a third-party entity that holds the document, wherein the contact information includes :a network address; an internet protocol (IP) address ;a host address; digital signatures; and a public key infrastructure information.
Furthermore, the cited art fails to specifically disclose generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the user validation includes UI indicators that specify portions of the data extraction prediction that are correct and incorrect, highlighting portions of the data extraction prediction with a mouse; and making a determination that the user validation indicates that the data extraction prediction is not correct based on the UI indicators. However, Bhandarkar et al discloses supervised or semi-supervised machine learning to generate the training dataset necessary to train the machine learned model. (0036) The model identifies a set of passages in the document to highlight wherein the highlighted text passage is designated by a visual indicator. (0035) Furthermore, the user is able to provide feedback regarding the correctness of each highlighted text passage. (0019, 022: the user may approve or reject one or more highlighted text passages identified and automatically highlighted) The user may delete one or more of the highlighted text passages 280 (for instance, in embodiments where the automatic highlight is incorrect). (0041) In other words, the user is able to indicate if a highlight text passage is correct or incorrect by maintaining the text passage being highlight or removing highlight from the highlight text passage. Thus, the presence of a text remains being highlighted indicates the portions of data are correct wherein if the text is unhighlighted indicates the portions are incorrect. This is a form of user validation. Furthermore, additional text passages can be highlighted. (0031, 0041) Bhandarkar discloses the user is able to manually highlight additional passages by selecting the text (user selection) (0031: The user may highlight and/or select candidate text passages for highlighting by receiving user input). While Bhandarkar et al does not explicitly state that a mouse was use to perform the manual highlighting of text by the user, a selection using an input device such as a mouse is well known in the art. The Examiner takes OFFICIAL NOTICE on this fact that it is typically known a mouse is used for selecting and highlighting text; therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified Bhandarkar’s ability to allow a user to provide an input to make a highlighting selection, such that the input selection could have been through an input device of a mouse as well known in the art. This combination would have provided the intrinsic advantage excels at selecting arbitrary or non-contiguous blocks of text with high precision. In addition, based on the unhighlighting or new highlighting by the user, the machine learning model is re-trained. (0041) Thus, Bhandarkar is able to determine if the user validation indicates identified text passage by the model is correct or incorrect if the text passage remains highlight (model incorrect identified the text passage when its highlight indicator is removed) Thus, Bhandarkar discloses in response to determining that a highlighted text passages is incorrect by its UI indicator (no longer highlight), the machine learning model for identifying relevant text (by highlighting) is re-trained.
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Bhandarkar since it would have provided the benefit of an improved, efficient, and more reliable document execution experience for the user.
Thus, in conjunction with Wilson et al, the combination teaches generate, by a user, a user validation using interaction capabilities of the GUI, wherein: the user validation includes UI indicators that specify portions of the data extraction prediction that are correct and incorrect, the interaction capabilities of the GUI comprise of: clicking a button; entering in information via a keyboard to provide feedback on portions of the data extraction prediction; and highlighting portions of the data extraction prediction with a mouse;
In addition, the cited art fails to specifically disclose generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the interaction capabilities of the GUI comprise of: checking a box. However, Tanikawa discloses generating, by a user, a user validation using interaction capabilities of the GUI, wherein: the interaction capabilities of the GUI comprise of: checking a box (0028: user checks a box to indicate data has been validated)
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Tanikawa since checking a box would have provided the intrinsic advantage of provides a clear, simple, visual way to confirm an item meets a specific condition or status, making data more reliable and easier to analyze.
Furthermore, the cited art fails to specifically discloses monitor, by a document data extractor, points in time on a prediction model update schedule for a model prediction update event, wherein the document data extractor is a physical hardware processor. However, Ellison discloses the data classification system 100 (e.g., classification computing entity 65) continuously retrains the machine learning models and applies the retrained models to new data assets, wherein the data classification system 100 can retrain the one or more machine learning models on a fixed schedule (e.g., hourly, daily, weekly, and/or the like). (0076-0077) 0028 discloses the classification computing entity 65 includes a processor. Thus, Ellison discloses a physical hardware processor that retrains the model on a fixed schedule, a form of monitoring by a “document data extractor”)
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Ellison since it would have provided the intrinsic advantage of improve accuracy over time and ensures the model periodically "re-learns" the current patterns.
Furthermore, as explained above, Wilson et al disclose continuously calculating an accuracy of the prediction model (0074: discloses training the one or more data extraction models, and validating them against additional sets of validation documents, may be repeated until a desired level of data extraction accuracy is achieved. Once the desired performance metrics for the trained extraction model(s) have been achieved, the one or more trained extraction models may be deployed; 0096: The performance of the term-based data extraction model may then be compared to a set of target performance metrics (e.g., annotation accuracy, etc.) at step 2218 to decide whether the model should be updated (e.g., further trained) or deployed for use. In other words, It appears that a form of a threshold was used since the model gets retrained until a desired level of data extraction accuracy is achieved. However, Wilson fails to specifically disclose making, based on the monitoring, a second determination that the accuracy is below a threshold identifying, based on the second determination, a prediction model update event. However, Matiukhov discloses a determination is made as to whether the predictions made by the model are below an accuracy threshold. In addition, Matiukhov discloses that if the model is below the explicit threshold, then the model is retrained and this process is repeated until it reaches the desired threshold. Thus, Matiukhov also discloses (0064-0065; Claim 8)
It would have been obvious to one of ordinary skill in the art before the effective filing date of Applicant’s invention to have modified the cited art with the cited feature(s) of Matiukhov since it would have provided the benefit an automated document data processing that is time-saving less prone to error. (0022)
As per dependent claim 18, Claim 18 recites similar limitation(s) as in Claim 4 and are rejected under similar rationale.
Response to Arguments
Applicant's arguments filed 3/18/26 have been fully considered but they are not persuasive.
On page 13, in regards to the 35 USC 112 rejection, Applicant’s amendment has overcome the pending issues in Claim 15 and 18 that were disclosed in the previous office action. However, the amendments made to the claims comprises new issues within the claims that were not previously presented. The Examiner respectfully points to the Applicant to the “Claim Rejections - 35 USC § 112” section of the office action above regarding the matter. Therefore, the 112 rejection(s) to the claims remain.
Applicant’s arguments with respect to claims 1, 8, 15 have been considered but are moot because the arguments do not apply to the new ground(s) of rejection(s) since the new ground(s) of rejection(s) was necessitated by Applicant's amendment.
Conclusion
If the Applicant chooses to amend the claims in future filings, the Examiner kindly states any new limitation(s) added to the claims must be described in the specification in such a way as to reasonably convey to one skilled in the relevant art in order to meet the written description requirement of 35 USC 112, first paragraph. To help expedite prosecution, promote compact prosecution and prevent a possible 112(a)/first paragraph rejection, the Examiner respectfully requests for each new limitation added to the claims in a future filing by the Applicant that the Applicant would cite the location within the specification showing support for that new limitation within the remarks. In addition, MPEP 2163.04(I)(B) states that a prima facie under 112(a)/first paragraph may be established if a claim has been added or amended, the support for the added limitation is not apparent, and applicant has not pointed out where added the limitation is supported.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DAVID FABER whose telephone number is (571)272-2751. The examiner can normally be reached Monday - Thursday.
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, Adam Queler can be reached at 5712724140. 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.
/ADAM M QUELER/ Supervisory Patent Examiner, Art Unit 2172
/D.F/ Examiner, Art Unit 2172
Prosecution Insights