DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This action is in response to communications filed December 12, 2025.
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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 December 12, 2025 has been entered.
Response to Arguments
Applicant's arguments filed November 26, 2025 regarding the rejection of claims 1-20 under 35 U.S.C 103 have been fully considered but they are moot in view of the new grounds of rejection.
Status of Claims
Claim 1-20 are pending, of which claims 1, 13, and 19 are in independent form. Claims 1-20 are rejected under 35 U.S.C. 103.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102 of this title, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries set forth in Graham v. John Deere Co., 383 U.S. 1, 148 USPQ 459 (1966), that are applied for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-5, 8, 12-17, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over Dixit et al. (US 2025/0103798) (hereinafter Dixit) in view of Chauhan et al. (US 2025/0061308) (hereinafter Chauhan), and in further view of Flores (US 2025/0200118) (hereinafter Flores) and Tsibulevskiy et al. (US 2022/0319219)(hereinafter Tsibulevskiy).
Regarding claim 1, Dixit teaches a method comprising: detecting, within an image, a first area that includes a first table (see Figs. 2-3, para [0020], para [0039], discloses recognizing characters and locations (first area) of data in images of invoices that detects the location and layout of data table (first table)); extracting, from the first area of the image, tabular data for the first table, the extracted tabular data comprising a first plurality of content items in the first table and first structural information for the first table (see Figs. 2-3, para [0022-0023], para [0026], discloses extracting tabular data comprising symbols and characters (first content items) in a first data table and columns and coordinates (first structural information) for the first data table).
Dixit does not explicitly teach generating a prompt that includes the first plurality of content items and the first structural information; providing the prompt as input to a first large language model; and responsive to providing the prompt as input to the first language model, generating, by the first language model, a first parsable representation of the first table, wherein the first parsable representation is in a first format and includes the first plurality of content items of the first table and the first structural information of the first table in the image.
Chauhan teaches responsive to providing the prompt as input to the first language model, generating, by the first large language model, a first parsable representation of the first table (see Figs. 2-3C, para [0045-0046], para [0049-0053], para [0063], discloses in response to labels, machine learning model generating HTML parsed data in cells (first parsable representation) of a HTML table (first table)), wherein the first parsable representation is in a first format and includes the first plurality of content items of the first table and the first structural information of the first table in the image (see Figs. 2-3C, para [0046], para [0049-0053], discloses HTML format (first format) and includes cell data of the HTML table (first table) and table data tags and classifies word or token in text as a type of named entity).
Dixit/Chauhan are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit to generate parsable representation from disclosure of Chauhan. The motivation to combine these arts is disclosed by Chauhan as “iteratively improve the prediction capabilities of the machine learning model, and may result in improved accuracy and consistency across different retailers, formats, and the like” (para [0075]) and generating parsable representation is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Dixit/Chauhan do not explicitly teach providing the prompt as input to a first large language model.
Flores teaches providing the prompt as input to a first large language model (see Fig. 3, para [0045], para [0064], discloses providing a prompt as input via Large Language Model, LLM).
Dixit/Chauhan/Flores are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan to provide prompt as input from disclosure of Flores. The motivation to combine these arts is disclosed by Flores as “increased impressions and improved user experiences with targeted/recommended content” (para [0074]) and providing prompt as input is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Dixit/Chauhan/Flores do not explicitly teach wherein the first structural information includes at least one coordinate value of a content item in the first plurality of content items and represents an area of the first table including the content item; generating a prompt that includes the first plurality of content items and the first structural information.
Tsibulevskiy teaches wherein the first structural information includes at least one coordinate value of a content item in the first plurality of content items and represents an area of the first table including the content item (see para [0045], para [0048], discloses coordinates (structural information) for regions (content items) of map of locations (first table)); generating a prompt that includes the first plurality of content items and the first structural information (see para [0045], para [0126], discloses generating figure (prompt) that include regions and coordinates).
Dixit/Chauhan/Flores/Tsibulevskiy are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores to include coordinate values of content items that represent an area of a first table from disclosure of Tsibulevskiy. The motivation to combine these arts is disclosed by Tsibulevskiy as “improve some accuracy of some annotations (or other forms of visual association)” (para [0147]) and including coordinate values of content items that represent an area of a first table is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Regarding claim 13, Dixit teaches a system comprising: one or more storage media storing instructions; and one or more processors configured to execute the instructions to cause the system to perform processing (see para [0005], discloses processor and memory storing instructions);comprising: detecting, within an image, a first area that includes a first table (see Figs. 2-3, para [0020], para [0039], discloses recognizing characters and locations (first area) of data in images of invoices that detects the location and layout of data table (first table)); extracting, from the first area of the image, tabular data for the first table, the extracted tabular data comprising a first plurality of content items in the first table and first structural information for the first table (see Figs. 2-3, para [0022-0023], para [0026], discloses extracting tabular data comprising symbols and characters (first content items) in a first data table and columns and coordinates (first structural information) for the first data table).
Dixit does not explicitly teach generating a prompt that includes the first plurality of content items and the first structural information; responsive to providing the prompt as input to the first language model, generating, by the first large language model, a first parsable representation of the first table, wherein the first parsable representation is in a first format and includes the first plurality of content items of the first table and the first structural information of the first table in the image.
Chauhan teaches responsive to providing the prompt as input to the first language model, generating, by the first large language model, a first parsable representation of the first table (see Figs. 2-3C, para [0045-0046], para [0049-0053], para [0063], discloses in response to labels, machine learning model generating HTML parsed data in cells (first parsable representation) of a HTML table (first table)), wherein the first parsable representation is in a first format and includes the first plurality of content items of the first table and the first structural information of the first table in the image (see Figs. 2-3C, para [0046], para [0049-0053], discloses HTML format (first format) and includes cell data of the HTML table (first table) and table data tags and classifies word or token in text as a type of named entity).
Dixit/Chauhan are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit to generate parsable representation from disclosure of Chauhan. The motivation to combine these arts is disclosed by Chauhan as “iteratively improve the prediction capabilities of the machine learning model, and may result in improved accuracy and consistency across different retailers, formats, and the like” (para [0075]) and generating parsable representation is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Dixit/Chauhan do not explicitly teach providing the prompt as input to a first large language model.
Flores teaches providing the prompt as input to a first large language model (see Fig. 3, para [0045], para [0064], discloses providing a prompt as input via Large Language Model, LLM).
Dixit/Chauhan/Flores are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan to provide prompt as input from disclosure of Flores. The motivation to combine these arts is disclosed by Flores as “increased impressions and improved user experiences with targeted/recommended content” (para [0074]) and providing prompt as input is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Dixit/Chauhan/Flores do not explicitly teach wherein the first structural information includes at least one coordinate value of a content item in the first plurality of content items and represents an area of the first table including the content item; generating a prompt that includes the first plurality of content items and the first structural information.
Tsibulevskiy teaches wherein the first structural information includes at least one coordinate value of a content item in the first plurality of content items and represents an area of the first table including the content item (see para [0045], para [0048], discloses coordinates (structural information) for regions (content items) of map of locations (first table)); generating a prompt that includes the first plurality of content items and the first structural information (see para [0045], para [0126], discloses generating figure (prompt) that include regions and coordinates).
Dixit/Chauhan/Flores/Tsibulevskiy are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores to include coordinate values of content items that represent an area of a first table from disclosure of Tsibulevskiy. The motivation to combine these arts is disclosed by Tsibulevskiy as “improve some accuracy of some annotations (or other forms of visual association)” (para [0147]) and including coordinate values of content items that represent an area of a first table is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Regarding claim 19, Dixit teaches one or more non-transitory computer-readable storage media storing instructions that, upon execution by one or more processors of a system (see Fig. 13, para [0108], discloses computer storage media and processor), cause the system to perform operations comprising: detecting, within an image, a first area that includes a first table (see Figs. 2-3, para [0020], para [0039], discloses recognizing characters and locations (first area) of data in images of invoices that detects the location and layout of data table (first table)); extracting, from the first area of the image, tabular data for the first table, the extracted tabular data comprising a first plurality of content items in the first table and first structural information for the first table (see Figs. 2-3, para [0022-0023], para [0026], discloses extracting tabular data comprising symbols and characters (first content items) in a first data table and columns and coordinates (first structural information) for the first data table).
Dixit does not explicitly teach generating a prompt that includes the first plurality of content items and the first structural information; responsive to providing the prompt as input to the first language model, generating, by the first large language model, a first parsable representation of the first table, wherein the first parsable representation is in a first format and includes the first plurality of content items of the first table and the first structural information of the first table in the image.
Chauhan teaches responsive to providing the prompt as input to the first language model, generating, by the first large language model, a first parsable representation of the first table (see Figs. 2-3C, para [0045-0046], para [0049-0053], para [0063], discloses in response to labels, machine learning model generating HTML parsed data in cells (first parsable representation) of a HTML table (first table)), wherein the first parsable representation is in a first format and includes the first plurality of content items of the first table and the first structural information of the first table in the image (see Figs. 2-3C, para [0046], para [0049-0053], discloses HTML format (first format) and includes cell data of the HTML table (first table) and table data tags and classifies word or token in text as a type of named entity).
Dixit/Chauhan are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit to generate parsable representation from disclosure of Chauhan. The motivation to combine these arts is disclosed by Chauhan as “iteratively improve the prediction capabilities of the machine learning model, and may result in improved accuracy and consistency across different retailers, formats, and the like” (para [0075]) and generating parsable representation is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Dixit/Chauhan do not explicitly teach providing the prompt as input to a first large language model.
Flores teaches providing the prompt as input to a first large language model (see Fig. 3, para [0045], para [0064], discloses providing a prompt as input via Large Language Model, LLM).
Dixit/Chauhan/Flores are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan to provide prompt as input from disclosure of Flores. The motivation to combine these arts is disclosed by Flores as “increased impressions and improved user experiences with targeted/recommended content” (para [0074]) and providing prompt as input is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Dixit/Chauhan/Flores do not explicitly teach wherein the first structural information includes at least one coordinate value of a content item in the first plurality of content items and represents an area of the first table including the content item; generating a prompt that includes the first plurality of content items and the first structural information.
Tsibulevskiy teaches wherein the first structural information includes at least one coordinate value of a content item in the first plurality of content items and represents an area of the first table including the content item (see para [0045], para [0048], discloses coordinates (structural information) for regions (content items) of map of locations (first table)); generating a prompt that includes the first plurality of content items and the first structural information (see para [0045], para [0126], discloses generating figure (prompt) that include regions and coordinates).
Dixit/Chauhan/Flores/Tsibulevskiy are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores to include coordinate values of content items that represent an area of a first table from disclosure of Tsibulevskiy. The motivation to combine these arts is disclosed by Tsibulevskiy as “improve some accuracy of some annotations (or other forms of visual association)” (para [0147]) and including coordinate values of content items that represent an area of a first table is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Regarding claim 20, Dixit/Chauhan/Flores/Tsibulevskiy teach a media of claim 19.
Dixit/Chauhan/Flores do not explicitly teach herein the prompt comprises an instruction for generating the first parsable representation.
Tsibulevskiy teaches herein the prompt comprises an instruction for generating the first parsable representation (see para [0126], para [0322], discloses figure comprises optical character recognition, OCR (parsable representation)).
Regarding claims 2, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit/Chauhan/Flores/Tsibulevskiy do not explicitly teach an example of task performance comprising: an example plurality of content items and an example structural information to input to the first large language model; and an example parsable representation generated based at least in part on the example plurality of content items and the example structural information.
Flores teaches an example plurality of content items and an example structural information to input to the first large language model (see Fig. 3, para [0047], para [0066], discloses discovery of information from network resources and extracting content from the network resource via LLM based on input).
Dixit/Chauhan/Flores do not explicitly teach parsable representation generated based at least in part on the example plurality of content items and the example structural information.
Tsibulevskiy teaches an example parsable representation generated based at least in part on the example plurality of content items and the example structural information (see para [0034], para [0126], discloses OCR based on regions of figures and coordinates).
Regarding claim 14, Dixit/Chauhan/Flores/Tsibulevskiy teach a system of claim 13.
Dixit further teaches wherein extracting the tabular data of the first table comprises: extracting each content item in the first plurality of content items from the first area of the image (see Fig. 2, para [0026], para [0039], discloses extracting characters and character locations).
Regarding claims 3 and 15, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1 and a system of claim 13.
Dixit further teaches wherein extracting the tabular data comprises: for each content item in the first plurality of content items, determining, location information for the content item, the location information for the content item indicative of a location of the content item within the image (see Figs. 3-4, para [0039], para [0045], discloses recognizing and detecting characters and locations of data in an invoice, indicating coordinates describing the location of a bounding box on the image); and wherein the first structural information for the first table includes the location information for the first plurality of content items represented (see Fig. 3, para [0039], para [0045], discloses indicating location and layout of invoice tables).
Regarding claim 4, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit further teaches at least one content item included in the first plurality of content items comprises at least one of: a word, a character, a symbol, or a value (see para [0045], discloses a character and value); and the first structural information includes at least two coordinate values identifying a bounding box around the at least one content item (see para [0045-0046], discloses a set of coordinates that describe a location of a bounding box on an image).
Regarding claim 5, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit does not explicitly teach wherein the first format is at least one of: a HyperText Markup Language (HTML) format, an Extensible Markup Language (XLM) format, or a comma separated values (CSV) format.
Chauhan teaches wherein the first format is at least one of: a HyperText Markup Language (HTML) format, an Extensible Markup Language (XLM) format, or a comma separated values (CSV) format (see para [0049], discloses HTML format).
Regarding claim 8, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit further teaches wherein the image includes a second table and the method further comprises: detecting, within the image, a second area that includes the second table (see Figs. 2-3, para [0020], para [0030], para [0039], discloses recognizing characters and locations (second area) of data in an image of invoice that detects the location and layout of a data table (second table)); extracting, from the second area of the image, second tabular data for the second table, the extracted second tabular data comprising a second plurality of content items in the second table and second structural information for the second table (see Figs. 2-3, para [0022-0023], para [0026], para [0039], discloses extracting tabular data comprising symbols and characters (second content items) in a second data table and columns and coordinates (second structural information) for the second data table); generating a second prompt that includes the second plurality of content items and the second structural information (see Fig. 4, para [0045-0046], discloses generating second label (second prompt) of a bounding box that includes text that describes text in bounding boxes that includes second coordinates and value string that is characters and symbols); providing the second prompt as input to the first language model (see Figs. 4-5, para [0047], para [0059], para [0063], discloses providing second label of bounding box determined by labeling model as input to text recognition model (first language model)).
Dixit does not explicitly teach responsive to providing the second prompt as input to the first language model, generating, by the first language model, a second parsable representation of the second table, wherein the second parsable representation is in a second format and includes the second plurality of content items of the second table, and the second structural information of the second table in the image.
Chauhan teaches responsive to providing the second prompt as input to the first language model, generating, by the first language model, a second parsable representation of the second table (see Figs. 2-4, para [0045-0046], para [0070], para [0082], discloses in response to labels, machine learning model generating HTML parsed data in cells (second parsable representation) of a JSON table (second table)), wherein the second parsable representation is in a second format and includes the second plurality of content items of the second table, and the second structural information of the second table in the image (see Figs. 2-4, para [0046], para [0070], para [0082], discloses JSON format (second format) and includes tokens of the JSON table (second table) and classifies word or token in text as a type of named entity).
Regarding claim 12, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit further teaches wherein the prompt further comprises: one or more examples including a first example, wherein the first example comprises a first portion and a second portion (see para [0046], discloses classification task that includes assigning labels to a bounding box (first and second portions)), the first portion identifying a plurality of example content items and corresponding example structural information for each example content item in the plurality of example content items (see para [0046], discloses labels for a column header of a label for table data), and the second portion identifying an example parsable representation corresponding to the first portion (see para [0046-0047], discloses bounding box coordinates that are labeled as table data).
Regarding claim 16, Dixit/Chauhan/Flores/Tsibulevskiy teach a system of claim 13.
Dixit/Chauhan does not explicitly teach wherein the processing further comprises: fine tuning the first large language model to enable the first large language model to extract tables from images; and wherein providing the prompt as input to the first large language model comprises providing the prompt to the first language model after performing the fine tuning.
Flores teaches wherein the processing further comprises: fine tuning the first large language model to enable the first large language model to extract tables from images (see Fig. 3, para [0070-0071], discloses fine-tuning model for extracting content based on altered input for subsequent search); and wherein providing the prompt as input to the first large language model comprises providing the prompt to the first language model after performing the fine tuning (see Fig. 3, para [0069],], discloses providing LLM prompt after determining fine-tuning requirement ).
Regarding claim 17, Dixit/Chauhan/Flores/Tsibulevskiy teach a system of claim 13.
Dixit does not explicitly teach wherein fine-tuning the first language model comprises performing at least one of: full fine-tuning or parameter efficient fine-tuning.
Chauhan teaches wherein fine-tuning the first large language model comprises performing at least one of: full fine-tuning or parameter efficient fine-tuning (see para [0044], para [0046], discloses machine learning model is fine-tuned on a specific task).
Claims 10 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Dixit et al. (US 2025/0103798) (hereinafter Dixit) in view of Chauhan et al. (US 2025/0061308) (hereinafter Chauhan) and Flores and Tsibulevskiy as applied to claim 1, and in further view of Thomas et al. (US 2024/0386058) (hereinafter Thomas).
Regarding claim 10, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit further teaches wherein the image includes a second table and the method further comprises: detecting, within the image, a second area that includes the second table (see Figs. 2-3, para [0020], para [0030], para [0039], discloses recognizing characters and locations (second area) of data in an image of invoice that detects the location and layout of a data table (second table)); extracting, from the second area of the image, second tabular data for the second table, the extracted second tabular data comprising a second plurality of content items in the second table and second structural information for the second table (see Figs. 2-3, para [0022-0023], para [0026], para [0039], discloses extracting tabular data comprising symbols and characters (second content items) in a second data table and columns and coordinates (second structural information) for the second data table); generating a second prompt that includes the second plurality of content items and the second structural information (see Fig. 4, para [0045-0046], discloses generating second label (second prompt) of a bounding box that includes text that describes text in bounding boxes that includes second coordinates and value string that is characters and symbols).
Dixit does not explicitly teach responsive to providing the second prompt as input to the second language model, generating, by the second language model, a second parsable representation of the second table, wherein the second parsable representation is in a second format and includes the second plurality of content items of the second table, and the second structural information of the second table in the image.
Chauhan teaches responsive to providing the second prompt as input to the second language model, generating, by the second language model, a second parsable representation of the second table (see Figs. 2-4, para [0045-0046], para [0049-0053], para [0063], discloses in response to labels, second machine learning model generating HTML parsed data in cells (second parsable representation) of a JSON table (second table)), wherein the second parsable representation is in a second format and includes the second plurality of content items of the second table, and the second structural information of the second table in the image (see Figs. 2-4, para [0046], para [0070], para [0082], discloses JSON format (second format) and includes tokens of the JSON table (second table) and classifies word or token in text as a type of named entity).
Dixit/Chauhan/Flores/Tsibulevskiy do not explicitly teach providing the second prompt as input to a second large language model different from the first language model.
Thomas teaches providing the second prompt as input to a second large language model different from the first language model (see Fig. 1, para [0035], para [0037], discloses providing second prompt to a second LLM, such as ChatGPT®, BERT, ERNIE, T5, XLNet, and the like).
Dixit/Chauhan/Flores/Tsibulevskiy/Thomas are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores/Tsibulevskiy to utilize a second LLM different from first LLM from disclosure of Thomas. The motivation to combine these arts is disclosed by Thomas as “reduced latency, and concomitant improvements to productivity costs” (para [0031]) and utilizing a second LLM different from first LLM is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Regarding claim 18, Dixit/Chauhan/Flores/Tsibulevskiy teach a system of claim 13.
Dixit/Chauhan/Flores/Tsibulevskiy do not explicitly teach wherein the first large language model is a decoder-only model or an encoder-decoder model.
Thomas teaches wherein the first large language model is a decoder-only model or an encoder-decoder model (see para [0035], discloses ChatGPT is decoder-only model).
Dixit/Chauhan/Flores/Tsibulevskiy/Thomas are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores/Tsibulevskiy to utilize a second LLM different from first LLM from disclosure of Thomas. The motivation to combine these arts is disclosed by Thomas as “reduced latency, and concomitant improvements to productivity costs” (para [0031]) and utilizing a second LLM different from first LLM is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Claims 6, 9, and 11 are rejected under 35 U.S.C. 103 as being unpatentable over Dixit et al. (US 2025/0103798) (hereinafter Dixit) in view of Chauhan et al. (US 2025/0061308) (hereinafter Chauhan) and Flores and Tsibulevskiy as applied to claim 1, and in further view of Berestovsky et al. (US 2023/0065915) (hereinafter Berestovsky).
Regarding claim 6, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit/Chauhan/Flores/Tsibulevskiy does not explicitly teach wherein the image is represented in a second format, different than the first format, the second format is at least one of: a Portable Document Format (PDF), a Joint Photographic Expert Group (JPEG) format, a Portable Network Graphics (PNG) format, Tag Image File Format (TIFF), or a Graphic Interchange Format (GIF).
Berestovsky teaches wherein the image is represented in a second format, different than the first format, the second format is at least one of: a Portable Document Format (PDF), a Joint Photographic Expert Group (JPEG) format, a Portable Network Graphics (PNG) format, Tag Image File Format (TIFF), or a Graphic Interchange Format (GIF) (see para [0042], discloses JPEG format (second format)).
Dixit/Chauhan/Flores/Tsibulevskiy/Berestovsky are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores/Tsibulevskiy to utilize a second format from disclosure of Berestovsky. The motivation to combine these arts is disclosed by Berestovsky as “improve how existing computers and machine learning models detect objects, such as tables” (para [0034]) and utilizing a second format is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Regarding claim 9, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit/Chauhan/Flores/Tsibulevskiy does not explicitly teach wherein the first table is in a first orientation in the image and the second table is in a second orientation in the image, the second orientation is different than the first orientation.
Berestovsky teaches wherein the first table is in a first orientation in the image and the second table is in a second orientation in the image, the second orientation is different than the first orientation (see Figs. 6B, Fig. 12, para [0051], para [0101, 0146], discloses first table orientation and a second table orientation for an image, utilizing deskew function).
Dixit/Chauhan/Flores/Tsibulevskiy/Berestovsky are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores/Tsibulevskiy to utilize a second format from disclosure of Berestovsky. The motivation to combine these arts is disclosed by Berestovsky as “improve how existing computers and machine learning models detect objects, such as tables” (para [0034]) and utilizing a second format is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Regarding claim 11, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit/Chauhan/Flores does not explicitly teach further comprising: creating a single joined parsable representation that includes both the first parsable representation of the first table and the second parsable representation of the second table.
Berestovsky teaches creating a single joined parsable representation that includes both the first parsable representation of the first table and the second parsable representation of the second table (see para [0033-0034], discloses extracting rows and columns from respective tables and aggregating row values based on a set of rules for automated functionality to associate values belonging to a same item).
Dixit/Chauhan/Flores/Tsibulevskiy/Berestovsky are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores/Tsibulevskiy to utilize a second format from disclosure of Berestovsky. The motivation to combine these arts is disclosed by Berestovsky as “improve how existing computers and machine learning models detect objects, such as tables” (para [0034]) and utilizing a second format is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Claims 7 is rejected under 35 U.S.C. 103 as being unpatentable over Dixit et al. (US 2025/0103798) (hereinafter Dixit) in view of Chauhan et al. (US 2025/0061308) (hereinafter Chauhan) and Flores and Tsibulevskiy as applied to claim 1, and in further view of Joy et al. (US 2022/0044134) (hereinafter Joy).
Regarding claim 7, Dixit/Chauhan/Flores/Tsibulevskiy teach a method of claim 1.
Dixit/Chauhan/Flores/Tsibulevskiy does not explicitly teach receiving a request from a requester to perform a table extraction operation on the image, the request further identifying the first format; and transmitting the first parsable representation of the first table to the requester.
Joy teaches receiving a request from a requester to perform a table extraction operation on the image, the request further identifying the first format (see para [0038], discloses receiving an input user query to perform optical character recognition, OCR (table extraction operation) on an image, identifying pdf format); and transmitting the first parsable representation of the first table to the requester (see Fig. 1, Fig. 4, para [0044], para [0051], discloses retrieving candidate answers from paragraph of table text).
Dixit/Chauhan/Flores/Tsibulevskiy/Joy are analogous arts as they are each from the same field of endeavor of database systems.
Before the effective filing date of the invention it would have been obvious to a person of ordinary skill in the art to modify the system of Dixit/Chauhan/Flores/Tsibulevskiy to identify a first format from disclosure of Joy. The motivation to combine these arts is disclosed by Joy as “extract relevant portions that enable answering the question” (para [0001]) and identifying a first format is well known to persons of ordinary skill in the art, and therefore one of ordinary skill would have good reason to pursue the known options within his or her technical grasp that would lead to anticipated success.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to COURTNEY HARMON whose telephone number is (571)270-5861. The examiner can normally be reached M-F 9am - 5pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Ann Lo can be reached at 571-272-9767. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/Courtney Harmon/Primary Examiner, Art Unit 2159