DETAILED ACTION
This action is in response to the original filing on December 27, 2023. Claims 1-20 are pending and have been considered below. Claims 1, 11, and 20 are independent claims.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Interpretation
The following is a quotation of 35 U.S.C. 112(f):
(f) Element in Claim for a Combination. – An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
The following is a quotation of pre-AIA 35 U.S.C. 112, sixth paragraph:
An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof.
Claims 11 (and its respective dependent claims 12-16) and 17 (and its respective dependent claims 18-19) in this application are given their broadest reasonable interpretation using the plain meaning of the claim language in light of the specification as it would be understood by one of ordinary skill in the art. The broadest reasonable interpretation of a claim element (also commonly referred to as a claim limitation) is limited by the description in the specification when 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is invoked.
As explained in MPEP § 2181, subsection I, claim limitations that meet the following three-prong test will be interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph:
(A) the claim limitation uses the term “means” or “step” or a term used as a substitute for “means” that is a generic placeholder (also called a nonce term or a non-structural term having no specific structural meaning) for performing the claimed function;
(B) the term “means” or “step” or the generic placeholder is modified by functional language, typically, but not always linked by the transition word “for” (e.g., “means for”) or another linking word or phrase, such as “configured to” or “so that”; and
(C) the term “means” or “step” or the generic placeholder is not modified by sufficient structure, material, or acts for performing the claimed function.
Use of the word “means” (or “step”) in a claim with functional language creates a rebuttable presumption that the claim limitation is to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites sufficient structure, material, or acts to entirely perform the recited function.
Absence of the word “means” (or “step”) in a claim creates a rebuttable presumption that the claim limitation is not to be treated in accordance with 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph. The presumption that the claim limitation is not interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, is rebutted when the claim limitation recites function without reciting sufficient structure, material or acts to entirely perform the recited function.
Claim limitations in this application that use the word “means” (or “step”) are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action. Conversely, claim limitations in this application that do not use the word “means” (or “step”) are not being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, except as otherwise indicated in an Office action.
Such claim limitations are:
“a data receiving subsystem configured to receive one or more electronic documents” in claim 11 (and its respective dependent claims 12-16). Here, “data receiving subsystem” is a generic placeholder (prong 1), modified by the function “configured to receive” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of receiving “one or more electronic documents.” The “data receiving subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“a data extraction subsystem configured to extract one or more data” in claim 11 (and its respective dependent claims 12-16). Here, “data extraction subsystem” is a generic placeholder (prong 1), modified by the function “configured to extract” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of extracting “one or more data.” The “data extracting subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“a linkage value determining subsystem configured to determine one or more first linkage values” in claim 11 (and its respective dependent claims 12-16). Here, “linkage value determining subsystem” is a generic placeholder (prong 1), modified by the function “configured to determine” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of determining “one or more first linkage values.” The “linkage value determining subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“a linkage key-value pair mapping subsystem configured to map one or more second linkage values” in claim 11 (and its respective dependent claims 12-16). Here, “linkage key-value pair mapping subsystem” is a generic placeholder (prong 1), modified by the function “configured to map” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of mapping “one or more second linkage values.” The “linkage key-value pair mapping subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“an end state mapping subsystem configured to map one or more first pairs” in claim 11 (and its respective dependent claims 12-16). Here, “end state mapping subsystem” is a generic placeholder (prong 1), modified by the function “configured to map” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of mapping “one or more first pairs.” The “end state mapping subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“a database updating subsystem configured to update the one or more second databases” in claim 11 (and its respective dependent claims 12-16). Here, “database updating subsystem” is a generic placeholder (prong 1), modified by the function “configured to update” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of updating “the one or more second databases.” The “database updating subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“an output subsystem configured to provide an output” in claim 11 (and its respective dependent claims 12-16). Here, “output subsystem” is a generic placeholder (prong 1), modified by the function “configured to provide” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of providing “an output.” The “output subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
“a training subsystem configured to train the machine learning model” in claim 17 (and its respective dependent claims 18-19). Here, “training subsystem” is a generic placeholder (prong 1), modified by the function “configured to train” (prong 2), and not modified by sufficient structure to perform the claimed function (prong 3). Specifically, the claimed “subsystem” is not sufficient structure to perform the function of training “the machine learning model.” The “training subsystem” is interpreted to mean “machine-readable instructions and/or source code stored in the memory… on any of the above-mentioned storage media and may be in communication with and executed by the one or more hardware processors” (¶67).
Because this/these claim limitation(s) is/are being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, it/they is/are being interpreted to cover the corresponding structure described in the specification as performing the claimed function, and equivalents thereof.
If applicant does not intend to have this/these limitation(s) interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph, applicant may: (1) amend the claim limitation(s) to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph (e.g., by reciting sufficient structure to perform the claimed function); or (2) present a sufficient showing that the claim limitation(s) recite(s) sufficient structure to perform the claimed function so as to avoid it/them being interpreted under 35 U.S.C. 112(f) or pre-AIA 35 U.S.C. 112, sixth paragraph.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 6-7 (and claim 7’s respective dependent claims 8-9), 10, 16-17 (and claim 17’s respective dependent claim 18), and 19 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claims 6-7 and 10 recite the limitation the one or more confidence scores. There is insufficient antecedent basis for this limitation in these claims. For examination purposes, “the one or more confidence scores” will be interpreted to mean “one or more confidence scores.”
Claim 6 recites the limitation the one or more linkage values. It is unclear which linkage values this limitation is referring to. For example, it could refer to the one or more first linkage values or the one or more second linkage values of claim 1. For examination purposes, “the one or more linkage values” will be interpreted to mean the one or more second linkage values of claim 1.
Claim 7 (and its respective dependent claims 8-9) recites the limitation the one or more second information and the optimum split. There is insufficient antecedent basis for these limitations in these claims. For examination purposes, “the one or more second information” will be interpreted to mean “one or more second information” and “the optimum split” will be interpreted to mean “an optimum split.”
Claims 16-17 (and claim 17’s respective dependent claim 18) and 19 are system claims that contain similar limitations to the methods of claims 6-7 (and claim 7’s respective dependent claim 8) and 10, respectively. Therefore, claims 16-19 are rejected under substantially the same rationale as claims 6-8 and 10, respectively.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding claim 1:
Step 1 – The claim is directed to a method: A machine-learning based (ML-based) computing method for automatic data mapping for one or more electronic documents…
Step 2A, Prong 1 – A judicial exception is recited in this claim as it recites mental processes (see MPEP 2106.04(a)(2)(III)):
determining… one or more first linkage values for the one or more first information associated with the one or more electronic documents by correlating each information of the one or more first information with one or more metadata extracted from the one or more first information… A human can reasonably determine “first linkage values” by correlating first information of an electronic document with metadata within the human mind or with the aid of a pen and paper, which is a mental process.
mapping… one or more second linkage values with one or more linkage keys, in one or more second databases, by: determining… whether the one or more first linkage values are matched with the one or more second linkage values in the one or more second databases, based on a fuzzy string matching technique… A human can reasonably perform “mapping” by “determining” whether linkage values are matched in a database using “a fuzzy string matching technique” within the human mind or with the aid of a pen and paper, which is a mental process.
determining… the one or more linkage keys by analyzing one or more columns in the one or more second databases where the one or more first linkage values are matched with the one or more second linkage values… A human can reasonably determine “linkage keys” by analyzing columns in a database within the human mind or with the aid of a pen and paper, which is a mental process.
Step 2A, Prong 2 – The following limitations are additional elements that fail to integrate the judicial exception into a practical application:
… by one or more hardware processors… processors used as a mere tool to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
receiving… one or more electronic documents from one or more first databases… receiving data from a database is mere data gathering (see MPEP 2106.05(g)).
extracting… one or more data comprising one or more first information from the one or more electronic documents based on an optical character recognition (OCR) conversion process… extracting is mere data gathering (see MPEP 2106.05(g)).
determining… based on a machine learning model… a machine learning model used as a mere tool to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
mapping… the one or more second linkage values with the one or more linkage keys… mapping a value to a key in a database is mere data gathering (see MPEP 2106.05(g)).
mapping… one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with one or more second pairs associated with at least one of: one or more end-state values and one or more end-state keys in the one or more second databases… mapping two key-value pairs in a database is mere data gathering (see MPEP 2106.05(g)).
updating… the one or more second databases by adjusting one or more payments associated with the one or more first information based on mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases… to update database values is mere data gathering (see MPEP 2106.05(g)).
and providing… an output of the updated one or more second databases with the adjusted one or more payments associated with the one or more first information to one or more first users on a user interface associated with one or more electronic devices… providing an output to a device for display is mere data outputting (see MPEP 2106.05(g)).
Step 2B – These elements are recited at such a high level of generality that they fail to integrate the abstract idea into a practical application, since they provide nothing more than mere instructions to implement an abstract idea on a generic computer (MPEP 2106.05(f)) or only amount to data gathering and outputting (MPEP 2106.05(g)) without significantly more. These limitations, taken either alone or in combination, fail to provide an inventive concept. Thus, the claim is not patent eligible.
Claims 2-10 recite limitations which further narrow the abstract idea of claim 1 by specifying more details of the mental processes that occur:
Regarding claim 2, this claim further limits the abstract idea of claim 1 to be based on a mental process: extracting… by analyzing a spatial layout of the one or more first information… wherein a human can reasonably perform “analyzing a spatial layout” within the human mind or with the aid of a pen and paper, which is a mental process. Furthermore, wherein extracting the one or more data comprising the one or more first information associated with the one or more electronic documents, comprises: extracting… the one or more first information… and extracting… the one or more metadata related to one or more coordinates associated with the one or more first information within the one or more electronic documents… is still mere data gathering (see MPEP 2106.05(g)). Furthermore, describing extracting, by the one or more hardware processors… is still implementing a judicial exception on a generic computer (MPEP 2106.05(f)). Furthermore, specifying first information in form of at least one of: one or more fields, one or more headers, one or more keywords, and one or more key numbers, within the one or more electronic documents and a spatial layout of the one or more first information, based on at least one of: one or more positions, one or more patterns, and one or more locations, of the one or more first information within the one or more electronic documents and the one or more metadata related to one or more coordinates associated with the one or more first information within the one or more electronic documents, wherein: the one or more first information comprise at least one of: one or more second users associated with the one or more electronic documents, one or more vendors linked to the one or more electronic documents, one or more invoice numbers, one or more invoice amounts corresponding to the one or more invoice numbers, one or more invoice dates, one or more payment details, one or more payment dates, one or more credit and one or more debit numbers, one or more check numbers, and one or more account numbers associated with the one or more electronic documents; and the one or more metadata related to the one or more coordinates associated with the one or more first information comprise at least one of: one or more top coordinates of one or more pages, one or more bottom coordinates of the one or more pages, one or more left coordinates of the one or more pages, one or more right coordinates of the one or more pages, one or more rows on the one or more pages, one or more sections on the one or more pages, one or more characters present in the one or more pages, one or more top coordinates of the one or more first information, one or more bottom coordinates of one or more first information, one or more left coordinates of the one or more first information, one or more right coordinates of the one or more first information, one or more distances of the one or more first information from a left side of the one or more pages, one or more distances of the one or more first information from a top of the one or more pages, one or more row numbers of the one or more first information in this manner does not overcome the rejection of claim 1 as modifying the “first information” and “metadata” does not make “analyzing” or “determining” to not be mental processes.
Regarding claim 3, this claim further limits the abstract idea of claim 1 to be based on mental processes: wherein determining… the one or more linkage values for the one or more first information associated with the one or more electronic documents, comprises: generating… one or more confidence scores for each information of the one or more first information in the one or more electronic documents; and labelling… the one or more first information to classify the one or more first information into at least one of: the one or more first linkage values and one or more second information, based on the one or more confidence scores generated for each information of the one or more first information by one or more predetermined threshold values, wherein the one or more second information are distinct from the one or more first information. A human can reasonably perform “generating… one or more confidence scores” for each information in an electronic document or “labelling… the one or more first information to classify the one or more first information” within the human mind or with the aid of a pen and paper. Furthermore, describing by the one or more hardware processors… is still implementing a judicial exception on a generic computer (MPEP 2106.05(f)).
Regarding claim 4, specifying wherein: the one or more first information are classified as the one or more second information when the one or more confidence scores for the one or more first information, are within the one or more predetermined threshold values; and the one or more first information are classified as the one or more first linkage values when the one or more confidence scores for the one or more first information exceed the one or more predetermined threshold value in this manner does not overcome the rejection of claim 3, as modifying “the one or more first information” does not make “determining” to not be a mental process.
Regarding claim 5, this claim further limits the abstract idea of claim 3 to be based on a mental process: generating… one or more data objects comprising the one or more confidence scores generated for each information of the one or more first information in the one or more electronic documents… A human can reasonably perform “generating… one or more data objects comprising the one or more confidence scores” within the human mind or with the aid of a pen and paper, which is a mental process. Furthermore, describing by the one or more hardware processors… is still implementing a judicial exception on a generic computer (MPEP 2106.05(f)). Furthermore, describing wherein when a single first information with the one or more confidence scores exceeds the one or more predetermined threshold values, the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases, and wherein when two or more first information with the one or more confidence scores exceed the one or more predetermined threshold values, the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases is still mere data gathering (see MPEP 2106.05(g)).
Regarding claim 6, describing when a single linkage key is determined in the one or more second databases, the one or more pairs of the single linkage key and the one or more linkage values comprising the one or more confidence scores are mapped with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases and when two or more linkage keys are determined in the one or more second databases, the one or more pairs of a linkage key and a corresponding linkage value comprising one or more optimum confidence scores are mapped with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases is still mere data gathering (see MPEP 2106.05(g)). Regarding claim 7, this claim further limits the abstract idea of claim 1 to be based on mental processes: selecting… one or more features vectors associated with at least one of: the one or more first information extracted from the one or more electronic documents and the metadata extracted from the one or more first information, wherein the machine learning model comprises a random forest based machine learning model and labelling… the one or more first information to classify the one or more first information into at least one of: the one or more first linkage values and the one or more second information, based on one or more pre-configured rules and parameters and segmenting… the one or more labelled datasets into at least one of: one or more training datasets and one or more validation datasets… to correlate the one or more feature vectors associated with at least one of: the one or more first information and the metadata, with the one or more first linkage values, based on one or more hyperparameters… and generating… the one or more confidence scores for the one or more first linkage values… A human can reasonably perform “selecting… one or more features vectors” and “labelling… the one or more first information” and “segmenting… the one or more labelled datasets” and correlating feature vectors associated with the one or more first information and metadata with the one or more first linkage values based on one or more hyperparameters, and “generating… the one or more confidence scores” within the human mind or with the aid of a pen and paper, which is a mental process. Furthermore, describing by the one or more hardware processors… and training… the machine learning model to correlate… and generating… based on the trained machine learning model is still implementing a judicial exception on a generic computer (MPEP 2106.05(f)). Furthermore, describing obtaining… one or more labelled datasets from the one or more first databases, wherein the one or more labelled datasets comprise the one or more first information extracted from the one or more electronic documents… and training… the machine learning model… is still mere data gathering (see MPEP 2106.05(g)). Furthermore, specifying hyperparameters comprise at least one of: max_depth, class_weight, n_estimators, criterion, min_samples_split, max_features, min_samples_leaf, and bootstrap, wherein the max_depth hyperparameter is configured to control an optimum depth of each decision tree in the random forest based machine learning model, wherein the class_weight hyperparameter is configured to adjust one or more weights of one or more classes in the random forest based machine learning model to control one or more class imbalance errors, wherein the n_estimators hyperparameter is configured to indicate a number of one or more decision trees to be included in the random forest based machine learning model, wherein the criterion hyperparameter is configured to determine quality of a split when the one or more decision trees are generated within the random forest based machine learning model, wherein the min_samples_split hyperparameter is configured to set a pre-determined number of one or more data points required in a node before the one or more data points split during a tree-building process, wherein the max_features hyperparameter is configured to determine an optimum number of features when the optimum split at each node in the random forest based machine learning model, wherein the min_samples_leaf hyperparameter is configured to indicate the pre-determined number of one or more data points required to generate a leaf node during the tree-building process, and wherein the bootstrap hyperparameter is configured to control whether the one or more data points are sampled when one or more individual decision trees are generated in the random forest based machine learning model in this manner does not overcome the rejection of claim 1 as modifying the “hyperparameters” does not make “determining” to not be a mental process.
Regarding claim 8, this claim further limits the abstract idea of claim 7 to be based on a mental process: further comprising validating… the machine learning model based on the one or more validation datasets, wherein validating the machine learning model comprises: determining… whether one or more metric scores attained by the trained machine learning model, exceeds one or more pre-determined threshold values, wherein the one or more metric scores are associated with one or more validation metrics comprising at least one of: precision metric, recall metric, F1-score metric, and confusion metric. A human can reasonably perform “validating… wherein validating the machine learning model comprises: determining… whether one or more metric scores… exceeds one or more pre-determined threshold values” within the human mind or with the aid of a pen and paper, which is a mental process. Furthermore, describing by the one or more hardware processors… is still implementing a judicial exception on a generic computer (MPEP 2106.05(f))
Regarding claim 9, this claim further limits the abstract idea of claim 8 to be based on a mental process: further comprising adjusting… the one or more hyperparameters to fine-tune the machine learning model based on one or more results of validation of the machine learning model. A human can reasonably perform “adjusting… the one or more hyperparameters… based on one or more results” within the human mind or with the aid of a pen and paper model, which is a mental process. Furthermore, describing by the one or more hardware processors… is still implementing a judicial exception on a generic computer (MPEP 2106.05(f)).
Regarding claim 10, this claim further limits the abstract idea of claim 7 to be based on a mental process: …to correlate the one or more feature vectors associated with at least one of: the one or more first information and the metadata, with the one or more first linkage values… and the one or more confidence scores are generated… and to determine the one or more first linkage values for the one or more first information associated with the one or more electronic documents. A human can reasonably correlate “the one or more feature vectors” and generate “the one or more confidence scores” and determine “the one or more first linkage values” for an electronic document within the human mind or with the aid of a pen and paper, which is a mental process. Furthermore, describing further comprising re-training… the machine learning model over a plurality of time intervals based on one or more training data, wherein re-training the machine learning model over the plurality of time intervals comprises: receiving… the one or more training data associated with at least one of: the one or more first information extracted from the one or more electronic documents and the metadata extracted from the one or more first information; adding… the one or more training data with the one or more training datasets to generate one or more updated training datasets; re-training… the machine learning model… is still mere data gathering (see MPEP 2106.05(g)). Furthermore, describing by the one or more hardware processors… and re-training… the machine learning model to correlate… and generated based on re-training the machine learning model… and executing… the re-trained machine learning model in a linkage value determining subsystem to determine… is still implementing a judicial exception on a generic computer (MPEP 2106.05(f)).
Regarding claim 11:
Step 1 – The claim is directed to a system: A machine learning based (ML-based) computing system for automatic data mapping for one or more electronic documents…
Step 2A, Prong 1 – A judicial exception is recited in this claim as it recites mental processes (see MPEP 2106.04(a)(2)(III)):
to determine one or more first linkage values for the one or more first information associated with the one or more electronic documents by correlating each information of the one or more first information with one or more metadata extracted from the one or more first information… A human can reasonably determine “first linkage values” by correlating first information of an electronic document with metadata within the human mind or with the aid of a pen and paper, which is a mental process.
to map one or more second linkage values with one or more linkage keys, in one or more second databases, by: determining whether the one or more first linkage values are matched with the one or more second linkage values in the one or more second databases, based on a fuzzy string matching technique… A human can reasonably map by “determining” whether linkage values are matched in a database using “a fuzzy string matching technique” within the human mind or with the aid of a pen and paper, which is a mental process.
determine the one or more linkage keys by analyzing one or more columns in the one or more second databases where the one or more first linkage values are matched with the one or more second linkage values… A human can reasonably determine “linkage keys” by analyzing columns in a database within the human mind or with the aid of a pen and paper, which is a mental process.
Step 2A, Prong 2 – The following limitations are additional elements that fail to integrate the judicial exception into a practical application:
one or more hardware processors… processors used as a mere tool to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
a memory coupled to the one or more hardware processors, wherein the memory comprises a plurality of subsystems in form of programmable instructions executable by the one or more hardware processors… a memory used as a mere tool to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
a data receiving subsystem configured to receive one or more electronic documents from one or more first databases… receiving data from a database is mere data gathering (see MPEP 2106.05(g)).
a data extraction subsystem configured to extract one or more data comprising one or more first information from the one or more electronic documents based on an optical character recognition (OCR) conversion process… extracting data is mere data gathering (see MPEP 2106.05(g)).
a linkage value determining subsystem configured to determine… as explained above in the interpretations under 112(f), using a processor to execute “a linkage value determining subsystem” or “machine-readable instructions and/or source code” to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
to determine… based on a machine learning model… a machine learning model used as a mere tool to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
a linkage key-value pair mapping subsystem configured to map… by: determining… as explained above in the interpretations under 112(f), using a processor to execute “a linkage key-value pair mapping subsystem” or “machine-readable instructions and/or source code” to apply an exception is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
mapping the one or more second linkage values with the one or more linkage keys… mapping a value to a key in a database is mere data gathering (see MPEP 2106.05(g)).
an end state mapping subsystem configured to map one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with one or more second pairs associated with at least one of: one or more end-state values and one or more end-state keys in the one or more second databases… mapping two key-value pairs in a database is mere data gathering (see MPEP 2106.05(g)).
a database updating subsystem configured to update the one or more second databases by adjusting one or more payments associated with the one or more first information based on mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases… to update database values by adjusting payments is mere data gathering (see MPEP 2106.05(g)).
an output subsystem configured to provide an output of the updated one or more second databases with the adjusted one or more payments associated with the one or more first information to one or more first users on a user interface associated with one or more electronic devices… providing an output to a device for display is mere data outputting (see MPEP 2106.05(g)).
Step 2B – These elements are recited at such a high level of generality that they fail to integrate the abstract idea into a practical application, since they provide nothing more than mere instructions to implement an abstract idea on a generic computer (MPEP 2106.05(f)) or only amount to data gathering and outputting (MPEP 2106.05(g)) without significantly more. These limitations, taken either alone or in combination, fail to provide an inventive concept. Thus, the claim is not patent eligible.
Regarding claim 17, describing a training subsystem configured to train the machine learning model, as explained above in the interpretations under 112(f), is using a processor to execute “a training subsystem” or “machine-readable instructions and/or source code” to apply an exception, which is a generic element for performing or applying the abstract idea using a generic computing environment (see MPEP 2106.05(f)).
Claims 12-16 and 18-19 are system claims that contain similar limitations to the methods of claims 2-6, 8 and 10, respectively. Therefore, claims 12-16 and 18-19 are rejected under substantially the same rationale as claims 2-6, 8 and 10, respectively.
Claim 20 is a computer-readable medium claim that contains similar limitations to the system of claim 11. Therefore, claim 20 is rejected under substantially the same rationale as claim 11.
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 1-2, 11-12, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Pedone et al. (US 20240338935 A1, hereinafter Pedone) in view of Agrawal et al. (US 20130312107 A1, hereinafter Agrawal).
Regarding claim 1:
Pedone teaches a machine-learning based (ML-based) computing method for automatic data mapping for one or more electronic documents (¶146 “Supervised natural language processing software then uses the best approximation mapping learned during training to analyze previously unseen input data to accurately predict the corresponding output,” ¶148 “Supervised learning software systems utilize neural network technology that includes, without limitation, Latent Semantic Analysis… Latent Semantic Analysis software processing techniques process a corpus of text data files to ascertain statistical co-occurrences of words that appear together which then yields insights into the subjects of those words and documents”), the ML-based computing method comprising: receiving, by one or more hardware processors (¶32 “The instructions, which execute via the processor of the computer or other programmable data processing apparatus, create mechanisms for implementing the functions specified in this specification and attached figures”), one or more electronic documents from one or more first databases (¶4 “The system can generate electronic transfer instruments using image data captured by an imaging device, such as a camera or scanner. The image data is processed using a variety of techniques to read and convert the information contained in the transfer instrument, which includes typed and handwritten text, into machine encoded content elements. The content elements, which can be individual characters, are grouped into tokens (e.g., words or groups of words) that represent transfer data,” Fig. 1 – 124, 134, 136, 146, ¶37 “The storage device 124 can store various other data items 134, including, without limitation, cached data, user files, pictures, audio and/or video recordings, files downloaded or received from other devices, and other data items preferred by the user,” ¶42 “The input and output system 136 may also include a camera 146, such as a digital camera”).
Pedone further teaches extracting, by the one or more hardware processors, one or more data comprising one or more first information from the one or more electronic documents based on an optical character recognition (OCR) conversion process (Fig. 11 – 252, 256-262, ¶114 “the Image Processing Module uses clustering analysis to perform OCR and identify characters. FIG. 11 depicts a flow diagram illustrating a process for identifying handwritten characters in an image. The transfer instrument includes a signature line 256, a line for source identification data 258, and a sequence data line 260 (i.e., a date), among other data fields. The images are subject to OCR processing at step 262 that reads and converts typed and/or printed text on the image 252 into machine-encoded text,” wherein text extracted from an image using OCR encompasses one or more data comprising one or more first information).
Pedone further teaches determining, by the one or more hardware processors, one or more first linkage values for the one or more first information associated with the one or more electronic documents by correlating each information of the one or more first information with one or more metadata extracted from the one or more first information, based on a machine learning model (¶127 “The segmentation analysis can also divide a transfer instrument or document by splitting the image data into text and non-text sections… A text section is a collection of human readable machine encoded content elements or characters that can be processed by an OCR system. Examples of a text section could be denoted by the text “pay to the order of” or a series of numerical characters representing routing data or a product identification number,” wherein “a product identification number” or other identifying information, for example, a “signature” or “source identification data” as explained above, encompasses one or more first linkage values for the one or more first information associated with the one or more electronic documents, ¶141 “Generating tokens can also be performed by detecting white spaces between machine encoded content elements… This may performed using a rules-based software engine alone or in combination with a hidden Markov model,” ¶142 “The segment analysis and token extraction can generate map index data that identifies the locations of segments and tokens with image data… each machine encoded content element… can be indexed in a sequence using pixel coordinates. Segments and tokens can be indexed according to a first coordinate index and an index length… transfer data for a target identification can have a location index coordinate and a length that maps the target identification data to a matrices of pixels,” wherein the “locations” of “pixel coordinates” in the image that are used to extract “tokens” of text which may contain, for example, identifying information, encompasses one or more metadata extracted from the one or more first information).
Pedone fails to teach mapping, by the one or more hardware processors, one or more second linkage values with one or more linkage keys, in one or more second databases by: determining, by the one or more hardware processors, whether the one or more first linkage values are matched with the one or more second linkage values in the one or more second databases, based on a fuzzy string matching technique. However, Agrawal, in the same field of endeavor, teaches this limitation (Abstract: “A computer receives an electronic document that includes a group of terms,” ¶21 “computer program instructions may be provided to a processor… such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified,” Fig. 1 – 135a-g, 140, 150, 155, ¶32 “sets of rules can be based on electronic dictionaries 135a-135g or regular expressions. A regular expression is program code that is utilized to search for strings of text such as specific terms or patterns of characters and/or numbers… a regular expression is used to identify the specific terms in the electronic document that match to structured data,” ¶33 “the information extraction program 140 uses a fuzzy matching technique. The fuzzy matching technique can identify the specific terms that match to a certain extent with attribute values within one of electronic dictionaries 135a-135g… the fuzzy matching technique can even identify the specific terms that match to a certain extent with attribute values within one of electronic dictionaries 135a-135g, wherein the specific terms are misspelled or have a typographical error,” wherein “specific terms in the electronic document” encompass the one or more first linkage values and “attribute values within one of electronic dictionaries” encompass one or more second linkage values, ¶34 “each of electronic dictionaries 135a-135… is an array of attribute values that are represented in an attribute column of database entity table 155 within MDM system 150. MDM system 150 is a centralized repository that contains structured data… each of electronic dictionaries 135a-135g and attribute values within electronic dictionaries 135a-135g are associated to an attribute column of database entity table 155,” ¶35 “each attribute column of database entity table 155 is associated to a column position index,” wherein an “a column position index” of the “database entity table” encompass one or more linkage keys, in one or more second databases).
Agrawal further teaches determining, by the one or more hardware processors, the one or more linkage keys by analyzing one or more columns in the one or more second databases where the one or more first linkage values are matched with the one or more second linkage values (Fig. 1 – 135a-g, 155, ¶35 “because each of electronic dictionaries 135a-135g and attribute values within electronic dictionaries 135a-135g are associated to an attribute column, each of electronic dictionaries 135a-135g and attribute values within electronic dictionaries 135a-135g are also associated to a column position index…The column position index is the location of the attribute column within database entity table 155… since the identified terms match to a certain extent with attribute values within one of electronic dictionaries 135a-135g, the identified terms are also each associated to an attribute column, a column position index”).
Agrawal further teaches mapping, by the one or more hardware processors, the one or more second linkage values with the one or more linkage keys (Fig. 1 – 140, Fig. 2 – 210b-216b, Fig. 3 – 302-320, ¶50 “information extraction program 140 determines column position indexes 210b-216b that are associated to each of specific terms 302-320, by utilizing functionality of the fuzzy matching. The column position indexes 210b-216b that are associated to each of specific terms 302-320 may be stored”).
Agrawal further teaches mapping, by the one or more hardware processors, one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with one or more second pairs associated with at least one of: one or more end-state values and one or more end-state keys in the one or more second databases (Fig. 1 – 130, 134, 150, 155, ¶38 “ Subsequent to document sensitivity computing program 130 performing fuzzy matching, document sensitivity computing program sends the identified terms to entity construction program 145 to perform record construction… After receiving the identified terms, entity construction program 145 sends a query that includes the identified terms, to MDM system 150… MDM system 150 determines one or more rows (i.e., entities) that may be associated to the identified terms, and returns the one or more rows to document sensitivity computing program 130,” wherein the “column position indexes” of the columns “associated to each of specific terms” determined through “fuzzy matching,” as explained above, encompass one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, ¶39 “document sensitivity computing program 130 performs additional processing to determine whether the one or more rows returned, if any, by MDM system 150 are actually associated to any of the identified terms… matching each of the identified terms to attribute values represented within the one or more rows in database entity table 155,” ¶40 “each row in database entity table 155 has a row position index… the row position index is the location of a row in database entity table 155… Document sensitivity computing program 130 includes functionality that can determine the row position index… actually associated with the identified terms,” wherein a “row position index” of the row “associated with the identified terms” encompasses one or more second pairs associated with at least one of: one or more end-state values and one or more end-state keys in the one or more second databases, when given its broadest reasonable interpretation of a second key-value pair that is mapped to the first key-value pair).
Regarding the limitation updating, by the one or more hardware processors, the one or more second databases by adjusting one or more payments associated with the one or more first information based on mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases, Pedone teaches updating, by the one or more hardware processors, the one or more second databases by adjusting one or more payments associated with the one or more first information based on a threshold (¶4 “The system can generate electronic transfer instruments using image data… The image data is processed using a variety of techniques to read and convert the information… which includes typed and handwritten text, into machine encoded content elements… which can be individual characters, are grouped into tokens (e.g., words or groups of words) that represent transfer data. The transfer data is used to implement the electronic transfer and can include information, such as a source identification… a target identification… or transfer value data,” Fig. 13, ¶227 “An example system for securing and validating electronic transfer instruments and detecting errors is shown in FIG. 13. The system includes a Secure Agent software engine that processes electronic transfer instruments, transfer activity data, system configuration data, and end user data to detect potential errors or instances of fraud… The Deposit Platform receives posting data from the Transfer Instrument Process Engine and completes the transfers by posting the transfers to a product account,” ¶238 “When the Secure Score meets the Secure Threshold, the electronic transfer instrument is sent to the Transfer Instrument Processing engine to generate the required posting data format for posting… the transfer is posted immediately such that the transfer data is added to the Transfer Activity Database for the particular target end user that received the transaction… the transfer value data is aggregated with the resource availability data in the Transfer Activity Databse. In the context of a check deposit, the result is that most or all of the funds would be immediately available for use,” ¶239 “Some or all of the transfer value data can be posted to the end user's resource availability data in the Transfer Activity Database… In the context of a check deposit, that means some of the funds from the check are made available,” Fig. 1 – 224, 234, ¶49 “The storage device 224 can store various other data 234, such as cached data, files for user accounts, user profiles, account balances, and transaction histories, files downloaded or received from other devices, and other data items required or related to the applications or programs,” wherein adding “transfer data” to “the Transfer Activity Database” encompasses updating… the one or more second databases by adjusting one or more payments). However, Pedone fails to teach updating… based on mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases.
Agrawal teaches the mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases (Fig. 1 – 130, 134, 135a-g, 150, 155, ¶¶33-35, 38-40, all as explained above).
Pedone further teaches and providing, by the one or more hardware processors, an output of the updated one or more second databases with the adjusted one or more payments associated with the one or more first information to one or more first users on a user interface associated with one or more electronic devices (Fig. 1 – 104, 106, 110, 140, 150, ¶40 “The integrated software applications also typically provide a graphical user interface (“GUI”) on the user computing device display screen 140 that allows the user 110 to utilize and interact with the user computing device,” ¶45 “The user computing device 104 & 106 further includes a communication interface 150. The communication interface 150 facilitates transactions with other devices and systems,” ¶215 “The system (i.e., a user computing device… can process source image data using a content recognition analysis to conduct an initial assessment regarding the quality of a digital image showing a transfer instrument… the initial assessment can determine readability of the transfer data contained on the transfer instrument… whether the product identification, the transfer value data, or other transfer data is readable such that it may be parsed or otherwise obtained and processed by the provider to execute the transfer. The initial quality assessment can be performed after a user captures all required images of the transfer instrument,” ¶216 “If the quality is confirmed, the user is prompted to confirm that the transaction is authorized and that electronic augmentations applied to the image are correct,” Fig. 13, ¶234 “In the Active Path, when the Secure Score meets or exceeds the Secure Threshold, the underlying transfer proceeds… The user computing device displays a transfer pass indicator or notification indicating that the transfer succeeded,” ¶235 “When a transfer is flagged, the system displays a transfer fail indicator or notification on the end user computing device indicating that the transfer was not completed”).
Pedone and Agrawal are analogous art to the claimed invention as both are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the fuzzy string matching and mapping of linkage values in the second database of Agrawal with the methodology of Pedone. The motivation to do so is to better associate data between a document and its identified user(s) (Agrawal, ¶26 “there is a need to generate a value associated with an electronic document based on information within the electronic document and the particular end-user that is attempting to access the electronic document. The value can be subsequently utilized to perform further action…”).
Regarding claim 2, Pedone in view of Agrawal teaches the machine-learning based (ML-based) computing method of claim 1 (and thus the rejection of claim 1 is incorporated).
Pedone further teaches wherein extracting the one or more data comprising the one or more first information associated with the one or more electronic documents, comprises: extracting, by the one or more hardware processors, the one or more first information in form of at least one of: one or more fields, one or more headers, one or more keywords, and one or more key numbers, within the one or more electronic documents by analyzing a spatial layout of the one or more first information, based on at least one of: one or more positions, one or more patterns, and one or more locations, of the one or more first information within the one or more electronic documents (¶135 “the first step in image processing is to process image data using a Fully Convolutional Neural Network that uses the images of the transfer instrument as inputs and yields as outputs a map of probabilities of attributes predicted for each pixel… each pixel is associated with probabilities that it belongs to a particular category of transfer data or transfer instrument segments that are identified,” ¶137 “the system can extract transfer data by grouping machine encoded content elements into tokens, such words or groups of words that comprise the transfer data… The machine encoded content elements are grouped into tokens, such as words, and groups of tokens are identified as one or more clusters… the string of machine encoded content elements… is recognized as a token “Patrick.” The token Patrick is identified as a name with semantic and heuristic processing techniques and grouped with additional tokens to identify a full cluster,” ¶138 “a string of machine encoded content elements and tokens “Deposit to Saving 1234” is identified as a cluster representing transfer instructions placed on a transfer instrument that denotes parameters for how an electronic transfer is conducted,” please note that first information in form of at least one of: one or more fields, one or more headers… and one or more key numbers has been interpreted to mean that the items in the list are disjunctive, therefore the limitation is interpreted to read “first information in form of: one or more fields OR one or more headers… OR one or more key numbers”).
Pedone further teaches and extracting, by the one or more hardware processors, the one or more metadata related to one or more coordinates associated with the one or more first information within the one or more electronic documents (¶142 as explained above with respect to claim 1), wherein: the one or more first information comprise at least one of: one or more second users associated with the one or more electronic documents, one or more vendors linked to the one or more electronic documents, one or more invoice numbers, one or more invoice amounts corresponding to the one or more invoice numbers, one or more invoice dates, one or more payment details, one or more payment dates, one or more credit and one or more debit numbers, one or more check numbers, and one or more account numbers associated with the one or more electronic documents (Fig. 11 – 256-260, ¶114 “The transfer instrument includes a signature line 256, a line for source identification data 258, and a sequence data line 260 (i.e., a date), among other data fields,” ¶115 “An example of handwritten text is shown as the transfer instruction “pay to the order of” positioned next to the source identification data line 256 of the transfer instrument. The OCR algorithm is programmed to identify the expected location of the certain characters in the image or other suitable typed text in the image 252, such as recognizing that sequencing data generally appears to the right of printed text stating “date” or that source identification data generally appears in the upper left portion of an image,” please note that comprise at least one of… has been interpreted to mean that the items in the subsequent list are disjunctive, as explained above).
Pedone further teaches and the one or more metadata related to the one or more coordinates associated with the one or more first information comprise at least one of: one or more top coordinates of one or more pages, one or more bottom coordinates of the one or more pages, one or more left coordinates of the one or more pages, one or more right coordinates of the one or more pages, one or more rows on the one or more pages, one or more sections on the one or more pages, one or more characters present in the one or more pages, one or more top coordinates of the one or more first information, one or more bottom coordinates of one or more first information, one or more left coordinates of the one or more first information, one or more right coordinates of the one or more first information, one or more distances of the one or more first information from a left side of the one or more pages, one or more distances of the one or more first information from a top of the one or more pages, one or more row numbers of the one or more first information (¶208 “To perform edge detection… an Image Processing Module software component first converts the image to black and white pixels with each pixel having position data (e.g., X-Y coordinates) and a brightness value indicating how light or how dark the pixel is to be displayed. The Image Processing Module analyzes adjacent rows and columns of pixels to determine abrupt changes in the brightness values that represent edges of the transfer instrument. The Image Processing Module can streamline the edge detection process by starting the analysis at locations where edges are expected, such as positions proximal to the visual guide locations or positions located a certain number of pixels away from the edge of the image,” please note that comprise at least one of… has been interpreted to mean that the items in the subsequent list are disjunctive, as explained above).
Regarding claim 11:
Pedone teaches a machine learning based (ML-based) computing system for automatic data mapping for one or more electronic documents (Fig. 1 – 100, ¶35 “a hardware system 100 configuration according to one embodiment”, ¶¶146, 148), the ML-based computing system comprising: one or more hardware processors (¶32).
Pedone further teaches a memory coupled to the one or more hardware processors, wherein the memory comprises a plurality of subsystems in form of programmable instructions executable by the one or more hardware processors (Fig. 1 – 120, 124-126, ¶37 “The storage device 124 includes at least one of a non-transitory storage medium for long-term, intermediate-term, and short-term storage of computer-readable instructions 126 for execution by the processor 120”), and wherein the plurality of subsystems comprises: a data receiving subsystem configured to receive (¶¶7-8 “a system for electronic transfer instrument security and error detection includes a computer with at least one processor and a memory device that stores data and executable code. The executable code causes the processor to transmit system configuration data to a network computer… The system activates a camera integrated or connected with the computer. The camera captures image data that comprises a transfer instrument image,” wherein “executable code” that causes the “system” to perform the function of “receiving” documents functions in substantially the same way as a data receiving subsystem as explained above in the interpretations under 112(f)) one or more electronic documents from one or more first databases (¶4, Fig. 1 – 124, 134, 136, 146, ¶¶37, 42).
Pedone further teaches a data extraction subsystem configured to extract one or more data comprising one or more first information from the one or more electronic documents based on an optical character recognition (OCR) conversion process (Fig. 11 – 252, 256-262, ¶114 as explained above with respect to claim 1; also note that an “Image Processing Module” functions in substantially the same way as a data extraction subsystem as explained above in the interpretations under 112(f)).
Pedone further teaches a linkage value determining subsystem configured to determine (¶140 “grouping machine encoded content elements can be performed by an extraction software module that is trained to label tokens and clusters according to classifications of transfer data, such as a transfer source identifier, a transfer target identification, a provider identifier, among other categories. The extraction software module can be implemented with a rule-based software technique, with probability models implemented by neural networks,” wherein “an extraction software module” functions in substantially the same way as a linkage value determining subsystem as explained above in the interpretations under 112(f)) one or more first linkage values for the one or more first information associated with the one or more electronic documents by correlating each information of the one or more first information with one or more metadata extracted from the one or more first information, based on a machine learning model (¶¶127, 141-142 all as explained above with respect to claim 1).
Pedone fails to teach a linkage key-value pair mapping subsystem configured to map one or more second linkage values with one or more linkage keys, in one or more second databases by: determining whether the one or more first linkage values are matched with the one or more second linkage values in the one or more second databases, based on a fuzzy string matching technique. However, Agrawal teaches this limitation (Abstract, ¶16 “aspects of the present invention may take the form of… an embodiment combining software and hardware aspects that may all generally be referred to herein as… ‘module’ or ‘system,’” ¶21, wherein “modules” or “computer program instructions” used to implement the described functions function in substantially the same way as a linkage key-value pair mapping subsystem, as explained above in the interpretations under 112(f), Fig. 1 – 135a-g, 140, 150, 155, ¶¶32-35 all as explained above with respect to claim 1).
Agrawal further teaches determining the one or more linkage keys by analyzing one or more columns in the one or more second databases where the one or more first linkage values are matched with the one or more second linkage values (Fig. 1 – 135a-g, 155, ¶35).
Agrawal further teaches and mapping the one or more second linkage values with the one or more linkage keys (Fig. 1 – 140, Fig. 2 – 210b-216b, Fig. 3 – 302-320, ¶50).
Agrawal further teaches an end state mapping subsystem configured to map (¶¶16, 21, wherein “modules” or “computer program instructions” used to implement the described functions function in substantially the same way as a linkage key-value pair mapping subsystem, as explained above in the interpretations under 112(f)) one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with one or more second pairs associated with at least one of: one or more end-state values and one or more end-state keys in the one or more second databases (Fig. 1 – 130, 134, 150, 155, ¶¶38-40 all as explained above with respect to claim 1).
Regarding the limitation a database updating subsystem configured to update the one or more second databases by adjusting one or more payments associated with the one or more first information based on mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases, Pedone teaches a database updating subsystem configured to update the one or more second databases by adjusting one or more payments associated with the one or more first information based on a threshold (¶4, Fig. 13, ¶¶227, 238-239, Fig. 1 – 224, 234, ¶49 all as explained above with respect to claim 1, ¶141 “a rules-based software engine,” ¶234 “the underlying transfer proceeds to the Deposit Platform after analysis by the Transfer Instrument Processing Engine” wherein “the Transfer Instrument Processing Engine” and “the Deposit Platform” function in substantially the same way as a database updating subsystem, as explained above in the interpretations under 112(f)). However, Pedone fails to teach to update… based on mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases.
Agrawal teaches the mapping of the one or more first pairs associated with at least one of: the one or more second linkage values and the one or more linkage keys, with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases (Fig. 1 – 130, 134, 135a-g, 150, 155, ¶¶33-35, 38-40, all as explained above with respect to claim 1).
Pedone further teaches and an output subsystem configured to provide an output of the updated one or more second databases with the adjusted one or more payments associated with the one or more first information to one or more first users on a user interface associated with one or more electronic devices (Fig. 1 – 104, 106, 110, 140, 150, ¶¶40, 45, 215-¶216, Fig. 13, ¶¶234-235, ¶162 “The Provider Interface Application can itself be comprised of one or more software services, software modules… that interface with the other hardware and software components integrated with a computing device,” ¶163 “the Provider Interface Application can include an Interface Service Module that generates graphical user interfaces (“GUIs”) displayed to users,” wherein “an Interface Service Module” functions in substantially the same way as an output subsystem, as explained above in the interpretations under 112(f)).
Pedone and Agrawal are analogous art to the claimed invention as both are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the fuzzy string matching and mapping of linkage values in the second database of Agrawal with the system and subsystems of Pedone. The motivation to do so is to better associate data between a document and its identified user(s) (Agrawal, ¶26 “there is a need to generate a value associated with an electronic document based on information within the electronic document and the particular end-user that is attempting to access the electronic document. The value can be subsequently utilized to perform further action…”).
Claim 12 is a system claim that contains similar limitations to the method of claim 2. Therefore, claim 12 is rejected under substantially the same rationale as claim 2.
Claim 20 is a computer-readable medium claim that contains similar limitations to the system of claim 11. Therefore, claim 20 is rejected under substantially the same rationale as claim 11.
Claims 3-5 and 13-15 are rejected under 35 U.S.C. 103 as being unpatentable over Pedone in view of Agrawal and further in view of Hertz et al. (US 20180082183 A1, hereinafter Hertz).
Regarding claim 3, Pedone in view of Agrawal teaches the machine-learning based (ML-based) computing method of claim 1 (and thus the rejection of claim 1 is incorporated).
Regarding the limitation wherein determining, by the machine learning model, the one or more first linkage values for the one or more first information associated with the one or more electronic documents, comprises: generating, by the one or more hardware processors, one or more confidence scores for each information of the one or more first information in the one or more electronic documents, Pedone teaches wherein determining, by the machine learning model, the one or more first linkage values for the one or more first information associated with the one or more electronic documents (Fig. 11 – 252-260, ¶¶114, 127 as explained above with respect to claim 1), comprises: generating, by the one or more hardware processors, one or more confidence scores for… the one or more first information in the one or more electronic documents (Fig. 13, ¶227 “An example system for securing and validating electronic transfer instruments and detecting errors is shown in FIG. 13. The system includes a Secure Agent software engine that processes electronic transfer instruments, transfer activity data, system configuration data, and end user data to detect potential errors or instances of fraud collectively referred to herein as “transfer tags” … The Secure Agent utilizes artificial intelligence and machine learning technology to determine a Secure Score that represents the likelihood an electronic transfer instrument is fraudulent or contains errors such that the underlying transfer should not be processed,” ¶246 “ As a further analysis for possible errors or instance of fraud, the Secure Agent can retrieve elements of end user data that can be verified against transfer data or augmentation data included as part of an electronic transfer instrument… After capturing an image of a transfer instrument, the system compares the user identification… to the transfer data on the transfer instrument, such as a user identification adjacent to a content element string “pay to the order of.” If the user identifications match, the system proceeds with creating an electronic transfer instrument. If the user identification do not match, the system can display a notification to the user indicating that the user identification could not be validated” wherein the “errors” are related to the extracted document text or the one or more first information in the one or more electronic documents, for example, mislabeled “user identification”). However, Pedone fails to teach one or more confidence scores for each information of the one or more first information…
Agrawal teaches a score for each information of a document (Fig. 1 – 130, Fig. 3 – 300, 302-320, ¶53 “To generate the value associated with electronic document 300, document sensitivity computing program 130 computes a first product of the column score and the row score if it is determined, for each of specific terms 302-320 to obtain sub-scores… if a row score is not determined for one of specific terms 302-320, then the sub-score for that identified term is merely the column score”).
Regarding the limitation and labelling, by the one or more hardware processors, the one or more first information to classify the one or more first information into at least one of: the one or more first linkage values and one or more second information, based on the one or more confidence scores generated for each information of the one or more first information by one or more predetermined threshold values, wherein the one or more second information are distinct from the one or more first information, Pedone teaches and labelling, by the one or more hardware processors, the one or more first information to classify the one or more first information into at least one of: fraudulent or not fraudulent (¶225 “Having extracted various elements of transfer data, the system can perform error checks by, for instance, comparing the legal transfer value data against the courtesy transfer value data to determine if the values are the same,” ¶236 “The Secure Threshold can represent a calculated likelihood that a transfer instrument is, or is not, fraudulent or contains errors”), based on the one or more confidence scores generated for… the one or more first information by one or more predetermined threshold values (¶236 “the Secure Threshold can be set to “99%,” which represents a 99% probability that the transfer does not include any transfer tags… a transaction is sent to the Deposit Platform only when the Secure Agent determines a Secure Score of 99% indicating at least a 99% probability that the underlying transaction is not fraudulent or erroneous,” ¶237 “The Secure Threshold can be predetermined by the provider… The provider can set the Secure Threshold to “95%” for transfer instruments having transfer value data of $1,000 or less and a “99%” for transfer instruments having a value of greater than $1,000… the provider demands a higher degree of confidence and accuracy for larger transfer amounts… for transfers of more than $1,000, the provider requires a 99% probability that the transfer instrument is not fraudulent or erroneous, but requires only a 95% degree of confidence for lower transfer amounts,” ¶238 “When the Secure Score meets the Secure Threshold, the electronic transfer instrument is sent to the Transfer Instrument Processing engine to generate the required posting data format for posting”). However, Pedone fails to teach and labelling… to classify the one or more first information into at least one of: the one or more first linkage values and one or more second information, based on the one or more confidence scores generated for each information of the one or more first information… wherein the one or more second information are distinct from the one or more first information.
Agrawal teaches scores generated for each information of the document (Fig. 1 – 130, Fig. 3 – 300, 302-320, ¶53). However, Agrawal fails to teach and labelling… to classify the one or more first information into at least one of: the one or more first linkage values and one or more second information… wherein the one or more second information are distinct from the one or more first information.
Hertz, in the same field of endeavor, teaches classifying document information into at least one of: the one or more first linkage values and one or more second information… wherein the one or more second information are distinct from the one or more first information (¶137 “Given a free text document, we first perform named entity recognition (NER) on the document to extract various types of entities, including companies, people, locations, events, etc.,” wherein identified “entities” encompass the one or more first linkage values, ¶140 “The core of this approach is a machine learning classifier that predicts the probability of a possible relationship for a given pair of identified entities… in a given sentence. This classifier uses a set of patterns to exclude noisy sentences, and then extracts a set of features from each sentence… features are based on various transformations and normalizations that are applied to each sentence (such as replacing identified entities by their type, omitting irrelevant sentence parts, etc.),” wherein discarded text that is distinct from “entities” or “features,” such as “noisy sentences” or “irrelevant sentence parts,” encompass one or more second information when given its broadest reasonable interpretation).
Pedone, Agrawal, and Hertz are analogous art to the claimed invention as all are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the term-level scoring and classification of Agrawal and Hertz, respectively, with the document-level scoring and classification of Pedone. The motivation to do so is to better associate data between a document and its identified user(s) (Agrawal, ¶26 “there is a need to generate a value associated with an electronic document based on information within the electronic document and the particular end-user that is attempting to access the electronic document. The value can be subsequently utilized to perform further action…”) and “to integrate… mined information from unstructured data with existing structured data and to ultimately generate insights for users based upon such integrated data” (Hertz, ¶207).
Regarding claim 4, Pedone in view of Agrawal and further in view of Hertz teaches the machine-learning based (ML-based) computing method of claim 3 (and thus the rejection of claim 3 is incorporated).
Regarding the limitation wherein: the one or more first information are classified as the one or more second information when the one or more confidence scores for the one or more first information, are within the one or more predetermined threshold values, Pedone teaches wherein: the one or more first information are classified as fraudulent when the one or more confidence scores for the one or more first information, are within the one or more predetermined threshold values (¶236 “The Secure Threshold can represent a calculated likelihood that a transfer instrument is, or is not, fraudulent or contains errors… the Secure Threshold can be set to “99%,” which represents a 99% probability that the transfer does not include any transfer tags… a transaction is sent to the Deposit Platform only when the Secure Agent determines a Secure Score of 99% indicating at least a 99% probability that the underlying transaction is not fraudulent or erroneous,” ¶237 “for transfers of more than $1,000, the provider requires a 99% probability that the transfer instrument is not fraudulent or erroneous, but requires only a 95% degree of confidence for lower transfer amounts,” wherein a document containing one or more first information classified as “fraudulent” will be labeled as such because its text “contains errors” and hence, its “Secure Score” does not exceed or is within the one or more predetermined threshold values or the “Secure Threshold”). However, Pedone fails to teach wherein: the one or more first information are classified as the one or more second information…
Hertz teaches classifying information as the one or more second information (¶¶137, 140 as explained above with respect to claim 3).
Pedone further teaches and the one or more first information are classified as the one or more first linkage values when the one or more confidence scores for the one or more first information exceed the one or more predetermined threshold values (Fig. 11 – 252-260, ¶¶114, 127 as explained above with respect to claim 1, ¶238 “When the Secure Score meets the Secure Threshold, the electronic transfer instrument is sent to the Transfer Instrument Processing engine to generate the required posting data format for posting… the transfer value data is aggregated with the resource availability data in the Transfer Activity Database. In the context of a check deposit, the result is that most or all of the funds would be immediately available for use,” wherein it is implicit that the one or more first information, or the identifying information extracted from the document or “electronic transfer instrument,” are classified as the one or more first linkage values, as identifying information is implied to not be erroneous or fraudulent when “the Secure Score meets the Secure Threshold” or when the one or more confidence scores for the one or more first information exceed the one or more predetermined threshold values, and hence the transaction is allowed to progress).
Pedone and Hertz are analogous art to the claimed invention as both are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the classification of Hertz with the score and threshold of Pedone. The motivation to do so is “to integrate… mined information from unstructured data with existing structured data and to ultimately generate insights for users based upon such integrated data” (Hertz, ¶207).
Regarding claim 5, Pedone in view of Agrawal and further in view of Hertz teaches the machine-learning based (ML-based) computing method of claim 3 (and thus the rejection of claim 3 is incorporated).
Regarding the limitation further comprising generating, by the one or more hardware processors, one or more data objects comprising the one or more confidence scores generated for each information of the one or more first information in the one or more electronic documents, Pedone teaches further comprising generating, by the one or more hardware processors, one or more data objects comprising the one or more confidence scores generated for… the one or more first information in the one or more electronic documents (¶8 “The transfer instrument is segmented into various components where transfer data can be found, such as a component for displaying a target identifier and another component where the value of the transfer is recorded,” ¶227 “The Secure Agent utilizes artificial intelligence and machine learning technology to determine a Secure Score that represents the likelihood an electronic transfer instrument is fraudulent or contains errors such that the underlying transfer should not be processed,” ¶241 “The Secure Agent determines the Secure Score utilizing a wide variety of data from the transfer instrument, the transfer data, the transfer activity database, the system configuration data, and the end user data… the input from the various data sources are fed through a neural network where weights are assigned to the inputs based on the relative significance of a given input in influencing the output Secure Score,” given its broadest reasonable interpretation, generating… one or more data objects comprising the one or more confidence scores is interpreted to mean that the one or more data objects generated can themselves be the generated one or more confidence scores). However, Pedone fails to teach confidence scores generated for each information of the one or more first information…
Agrawal teaches scores generated for each information of the document (Fig. 1 – 130, Fig. 3 – 300, 302-320, ¶53).
Regarding the limitation wherein when a single first information with the one or more confidence scores exceeds the one or more predetermined threshold values, the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases, Pedone teaches wherein when a document with the one or more confidence scores exceeds the one or more predetermined threshold values, complete a transaction (¶238 “When the Secure Score meets the Secure Threshold, the electronic transfer instrument is sent to the Transfer Instrument Processing engine to generate the required posting data format for posting… In the context of a check deposit, the result is that most or all of the funds would be immediately available for use”). However, Pedone fails to teach a single first information with the one or more confidence scores and the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases.
Agrawal teaches a single first information with a score (Fig. 3 – 300, 302-320, ¶53 “the sub-score for that identified term,” wherein a single “term” extracted from a text document encompasses a single first information) and the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases (Fig. 1 – 135a-g, 140, 150, 155, ¶¶32-35, all as explained above with respect to claim 1).
Regarding the limitation and wherein when two or more first information with the one or more confidence scores exceed the one or more predetermined threshold values, the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases, Pedone teaches wherein when a document with the one or more confidence scores exceed the one or more predetermined threshold values, complete a transaction (¶238). However, Pedone fails to teach two or more first information with the one or more confidence scores and the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases.
Agrawal teaches two or more first information each with a score (Fig. 1 – 130, Fig. 3 – 300, 302-320, ¶53 “document sensitivity computing program 130 computes a first product of the column score and the row score if it is determined, for each of specific terms 302-320 to obtain sub-scores,” Fig. 3 depicts a plurality of terms or two or more first information extracted from a document) and the one or more second linkage values are mapped with the one or more linkage keys in the one or more second databases (Fig. 1 – 135a-g, 140, 150, 155, ¶¶32-35, all as explained above with respect to claim 1).
Pedone and Agrawal are analogous art to the claimed invention as both are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the term-level scores and database mapping of Agrawal with the document-level score and threshold-based condition of Pedone. The motivation to do so is to better associate data between a document and its identified user(s) (Agrawal, ¶26 “there is a need to generate a value associated with an electronic document based on information within the electronic document and the particular end-user that is attempting to access the electronic document. The value can be subsequently utilized to perform further action…”).
Claims 13-15 are system claims that contain similar limitations to the methods of claims 3-5, respectively. Therefore, claims 13-15 are rejected under substantially the same rationale as claims 3-5, respectively.
Claims 6 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Pedone in view of Agrawal and further in view of Anthony et al. (US 11593555 B1, hereinafter Anthony).
Regarding claim 6, Pedone in view of Agrawal teaches the machine-learning based (ML-based) computing method of claim 1 (and thus the rejection of claim 1 is incorporated).
Agrawal teaches wherein: when a single linkage key is determined in the one or more second databases, the one or more pairs of the single linkage key and the one or more linkage values comprising the one or more confidence scores (Fig. 1 – 135a-g, 140, 155, ¶35 “each attribute column of database entity table 155 is associated to a column position index and a column score… The column position index is the location of the attribute column within database entity table 155… the column score represents a level of importance, in regard to data security, that an enterprise places on the identified terms,” ¶36 “information extraction program 140 uses a fuzzy matching technique to perform a comparison of electronic dictionaries 135a-135g with the specific terms extracted from the electronic document. The comparison can identify the specific terms that match to a certain extent with attribute values within one of electronic dictionaries 135a-135g, and also identify the column position index and the column score associated to each of the identified terms”) are mapped with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases (Fig. 1 – 130, 145, 150, 155, ¶38 “Record construction refers to identifying one or more rows (i.e., entities) in database entity table 155 to which the identified terms are potentially associated. After receiving the identified terms, entity construction program 145 sends a query that includes the identified terms, to MDM system 150… MDM system 150 determines one or more rows (i.e., entities) that may be associated to the identified terms,” ¶39 “document sensitivity computing program 130 performs additional processing to determine whether the one or more rows returned, if any, by MDM system 150 are actually associated to any of the identified terms. The additional processing includes matching each of the identified terms to attribute values represented within the one or more rows in database entity table 155,” ¶40 “each row in database entity table 155 has a row position index and a row score,” Fig. 4A – 405 and 420).
Regarding the limitation and when two or more linkage keys are determined in the one or more second databases, the one or more pairs of a linkage key and a corresponding linkage value comprising one or more optimum confidence scores are mapped with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases, Agrawal teaches and when two or more linkage keys are determined in the one or more second databases (Fig. 1 – 130, 155, Fig. 2 – 210a-216a and 210b-216b, Fig. 3 – 300, 302-320, ¶47 “Document sensitivity computing program 130 receives electronic document 300… document sensitivity computing program 130 identifies attribute values in each attribute column 210a-216a of database entity table 155,” ¶50 “specific terms 302-320 are each associated with one of attribute columns 210a-216a, they also are each associated with one of respective column position indexes 210b-216b,” Fig. 2 depicts a plurality of identified columns, each with their own indexes, or when two or more linkage keys are determined), the one or more pairs of a linkage key and a corresponding linkage value comprising one or more… confidence scores (Fig. 1 – 135a-g, 140, 155, ¶¶35-36) are mapped with the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys in the one or more second databases (Fig. 1 – 130, 145, 150, 155, Fig. 2 – 155 depicts wherein a plurality of columns or the one or more pairs of a linkage key and a corresponding linkage value are mapped to a row, or the one or more second pairs associated with at least one of: the one or more end-state values and the one or more end-state keys, Fig. 4A – 405 and 420, ¶¶38-40). However, Agrawal fails to teach corresponding linkage value comprising one or more optimum confidence scores…
Anthony, in the same field of endeavor, teaches a key-value comprising an optimum score (Fig. 2 – 210, 212, Col. 4, Lines 61-66 “the score analysis module 212 selects the key-value pair with the best (e.g., highest) accuracy value determined by the confidence score model 210. The score analysis module 212 is also configured to identify the key-value pair with the highest match confidence”).
Pedone, Agrawal, and Anthony are analogous art to the claimed invention as all are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the key-value pair mapping of Agrawal and the optimum scoring of Anthony with the methodology of Pedone. The motivation to do so is to better associate data between a document and its identified user(s) (Agrawal, ¶26 “there is a need to generate a value associated with an electronic document based on information within the electronic document and the particular end-user that is attempting to access the electronic document. The value can be subsequently utilized to perform further action…”) and to “utilize various rule-based analytical techniques to predict the final value for [a] field based on… duplicate fields” (Anthony, Col. 2, Lines 41-43).
Claim 16 is a system claim that contains similar limitations to the method of claim 6. Therefore, claim 16 is rejected under substantially the same rationale as claim 6.
Claims 7-10 and 17-19 are rejected under 35 U.S.C. 103 as being unpatentable over Pedone in view of Agrawal and further in view of Hertz, and further in view of Tanniru et al. (US 20210295103 A1, hereinafter Tanniru).
Regarding claim 7, Pedone in view of Agrawal teaches the machine-learning based (ML-based) computing method of claim 1 (and thus the rejection of claim 1 is incorporated).
Pedone further teaches further comprising training, by the one or more hardware processors, the machine learning model, by: obtaining, by the one or more hardware processors, one or more labelled datasets from the one or more first databases, wherein the one or more labelled datasets comprise the one or more first information extracted from the one or more electronic documents (¶8 “The camera captures image data that comprises a transfer instrument image. The image data is read and converted to machine encoded content elements, such as handwritten or typed text characters, to identify text on the transfer instrument image,” Fig. 1 – 124, 134, 136, 146, ¶37 “The storage device 124 can store various other data items 134, including, without limitation, cached data, user files, pictures, audio and/or video recordings, files downloaded or received from other devices, and other data items preferred by the user,” ¶42 “The input and output system 136 may also include a camera 146, such as a digital camera,” Fig. 13, ¶227 “An example system for securing and validating electronic transfer instruments… to detect potential errors or instances of fraud collectively referred to herein as ‘transfer tags,’” ¶251 “The historical secure agent database record can include the results of whether a transfer tag was detected… The labeling analysis determines whether transfer tags were present in a particular transaction but not caught by the Secure Agent. The historical labeling data is stored to the historical resource initialization database record to create a training data set,” wherein the “training data set,” or one or more labelled datasets, can only be obtained after labeling data received from the one or more first databases or “the storage device,” which comprises the one or more first information extracted from the one or more electronic documents or text extracted from the “transfer instrument” captured by a “camera”).
Pedone further teaches selecting, by the one or more hardware processors, one or more features vectors associated with at least one of: the one or more first information extracted from the one or more electronic documents and the metadata extracted from the one or more first information (¶126 “segment analysis segments or divides a transfer instrument into logical sections or components. The segmentation can be based on transfer structure, such as lines between text, segment titles, line breaks, indentations, or combinations of such features,” ¶127 “segmentation analysis can also divide a transfer instrument or document by splitting the image data into text and non-text sections,” ¶241 “When the Secure Agent is implemented by neural networking technology, the input from the various data sources are fed through a neural network where weights are assigned to the inputs based on the relative significance of a given input in influencing the output Secure Score,” ¶247 “any flagged OCR data is provided as an input to the neural network with a particular weight that varies depending on the significance of the input. If a misplaced provider logo image is known, based on historical data to correspond to higher likelihoods of fraud, then that input is assigned a higher weight for processing by the neural network,” ¶248 “the routing data is used to determine the source provider identification… The source provider identification is compared against a list of providers that demonstrate higher instance of fraudulent transfers. If there is a match… then the routing data is assigned a higher weight,” wherein a series of “weights” that are assigned to various extracted “inputs” or features, such as a “provider logo” or “routing data,” functions in substantially the same way as one or more feature vectors associated with at least one of: the one or more first information extracted from the one or more electronic documents and the metadata extracted, when given its broadest reasonable interpretation of a numerical representation of text for machine learning processing), wherein the machine learning model comprises a random forest based machine learning model (¶88 “the machine learning program may include… Random Forest (“RF”)”).
Regarding the limitation labelling, by the one or more hardware processors, the one or more first information to classify the one or more first information into at least one of: the one or more first linkage values and the one or more second information, based on one or more pre-configured rules and parameters, Pedone teaches labelling, by the one or more hardware processors, the one or more first information to classify the one or more first information into at least one of: fraudulent or not fraudulent, based on one or more pre-configured rules and parameters (¶¶225, 236-238 as explained above with respect to claim 3). However, Pedone fails to teach to classify the one or more first information into at least one of: the one or more first linkage values and the one or more second information…
Hertz teaches classifying document text into at least one of: the one or more first linkage values and one or more second information (¶¶137, 140 as explained above with respect to claim 3).
Pedone further teaches segmenting, by the one or more hardware processors, the one or more labelled datasets into at least one of: one or more training datasets and one or more validation datasets (¶251-252 “The labeling analysis determines whether transfer tags were present in a particular transaction but not caught by the Secure Agent. The historical labeling data is stored to the historical resource initialization database record to create a training data set… The training data set is input to neural network software applications and machines that implement the Secure Agent. The outputs of the neural network software applications are evaluated to determine whether the outputs match the historical labeling data within pre-defined error rates,” wherein “historical labeling data” that is used as both “a training data set” and to evaluate “whether the outputs match” implies segmenting… the one or more labelled datasets into… one or more training datasets and one or more validation datasets).
Regarding the limitation training, by the one or more hardware processors, the machine learning model to correlate the one or more feature vectors associated with at least one of: the one or more first information and the metadata, with the one or more first linkage values, based on one or more hyperparameters, wherein the one or more hyperparameters comprise at least one of: max_depth, class_weight, n_estimators, criterion, min_samples_split, max_features, min_samples_leaf, and bootstrap, wherein the max_depth hyperparameter is configured to control an optimum depth of each decision tree in the random forest based machine learning model, wherein the class_weight hyperparameter is configured to adjust one or more weights of one or more classes in the random forest based machine learning model to control one or more class imbalance errors, wherein the n_estimators hyperparameter is configured to indicate a number of one or more decision trees to be included in the random forest based machine learning model, wherein the criterion hyperparameter is configured to determine quality of a split when the one or more decision trees are generated within the random forest based machine learning model, wherein the min_samples_split hyperparameter is configured to set a pre-determined number of one or more data points required in a node before the one or more data points split during a tree-building process, wherein the max_features hyperparameter is configured to determine an optimum number of features when the optimum split at each node in the random forest based machine learning model, wherein the min_samples_leaf hyperparameter is configured to indicate the pre-determined number of one or more data points required to generate a leaf node during the tree-building process, and wherein the bootstrap hyperparameter is configured to control whether the one or more data points are sampled when one or more individual decision trees are generated in the random forest based machine learning model, Pedone teaches limitation training, by the one or more hardware processors, the machine learning model to correlate the one or more feature vectors associated with at least one of: the one or more first information and the metadata, with the one or more first linkage values (¶127 “segmentation analysis can also divide a transfer instrument or document by splitting the image data into text and non-text sections… Examples of a non-text section could include a provider logo or symbol… A text section is a collection of human readable machine encoded content elements or characters that can be processed by an OCR system,” ¶241 “the Secure Agent is implemented by neural networking technology, the input from the various data sources are fed through a neural network where weights are assigned to the inputs based on the relative significance of a given input in influencing the output Secure Score,” ¶247 “any flagged OCR data is provided as an input to the neural network with a particular weight that varies depending on the significance of the input. If a misplaced provider logo image is known, based on historical data to correspond to higher likelihoods of fraud, then that input is assigned a higher weight for processing by the neural network,” ¶251 “The parameters (e.g., node weights and formulas) of the Secure Agent are continuously adjusted to improve the accuracy of transfer tag detection,” ¶252 “training data set is input to neural network software applications and machines that implement the Secure Agent… The weighting coefficients of the neural network software applications are adjusted to reduce the error rates, and the process is run iteratively to train the neural network,” wherein “weights” assigned to various inputs, or the one or more feature vectors associated with at least one of: the one or more first information and the metadata, as explained above, are correlated to the one or more first linkage values, or identifying information extracted from documents that may be fraudulent or erroneous, by the machine learning model or “Secure Agent” trained to “improve the accuracy of transfer tag detection” or fraud/error detection). However, Pedone fails to teach based on one or more hyperparameters, wherein the one or more hyperparameters comprise at least one of: max_depth, class_weight…
Tanniru, in the same field of endeavor, teaches tuning a random forest model based on one or more hyperparameters, wherein the one or more hyperparameters comprise at least one of: max_depth… n_estimators… wherein the max_depth hyperparameter is configured to control an optimum depth of each decision tree in the random forest based machine learning model… wherein the n_estimators hyperparameter is configured to indicate a number of one or more decision trees to be included in the random forest based machine learning model… (Fig. 2 –235, 240, ¶67 “the machine learning system may use one or more hyperparameter sets 240 to tune the machine learning model… Example hyperparameters for a decision tree algorithm include a tree ensemble technique to be applied (e.g., bagging, boosting, a random forest algorithm… and/or the like), a number of features to evaluate, a number of observations to use, a maximum depth of each decision tree (e.g., a number of branches permitted for the decision tree), a number of decision trees to include in a random forest algorithm, and/or the like,” please note that hyperparameters comprise at least one of: max_depth, class_weight… and bootstrap has been interpreted to mean that the items in the list are disjunctive, therefore the limitation is interpreted to read “hyperparameters comprise: max_depth OR class_weight… OR bootstrap”).
Pedone further teaches and generating, by the one or more hardware processors, the one or more confidence scores for the one or more first linkage values, based on the trained machine learning model (¶198 “transfer data… can include… source identification data that names or identifies the source of the resource transfer… sequencing data, such as a date… target identification data that names or identifies the user receiving the data… transfer value data… signature data,” ¶225 “Having extracted various elements of transfer data, the system can perform error checks,” ¶227 “The Secure Agent utilizes artificial intelligence and machine learning technology to determine a Secure Score that represents the likelihood an electronic transfer instrument is fraudulent or contains errors such that the underlying transfer should not be processed”).
Pedone, Hertz, and Tanniru are analogous art to the claimed invention as all are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the classification of Hertz and the random forest hyperparameters of Tanniru with the methodology of Pedone. The motivation to do so is “to integrate… mined information from unstructured data with existing structured data and to ultimately generate insights for users based upon such integrated data” (Hertz, ¶207) and to “improve speed and efficiency of the [image-based document processing] and conserve computing resources (e.g., processing resources, memory resources, communication resources, and/or the like), networking resources, and/or the like” (Tanniru, ¶56).
Regarding claim 8, Pedone in view of Agrawal and further in view of Hertz, and further in view of Tanniru teaches the machine-learning based (ML-based) computing method of claim 7 (and thus the rejection of claim 7 is incorporated).
Regarding the limitation further comprising validating, by the one or more hardware processors, the machine learning model based on the one or more validation datasets, wherein validating the machine learning model comprises: determining, by the one or more hardware processors, whether one or more metric scores attained by the trained machine learning model, exceeds one or more pre-determined threshold values, wherein the one or more metric scores are associated with one or more validation metrics comprising at least one of: precision metric, recall metric, F1-score metric, and confusion metric, Pedone teaches validating, by the one or more hardware processors, the machine learning model based on the one or more validation datasets (¶252 “The outputs of the neural network software applications are evaluated to determine whether the outputs match the historical labeling data within pre-defined error rates”), wherein validating the machine learning model comprises: determining, by the one or more hardware processors, whether one or more metric scores attained by the trained machine learning model, exceeds one or more pre-determined threshold values, wherein the one or more metric scores are associated with one or more validation metrics comprising at least one of: success rate (¶232 “The provider can establish an active path threshold such that when the Secure Agent can detect nine (9) out of the ten (10) known transfer tags, the Secure Agent is deemed to be performing sufficiently,” ¶233 “The active path threshold can also… result in detecting seven (7) of the ten (10) known transfer tags in the set of one-hundred (100) transfer instruments… when the Secure Agent can also detect seven out of ten transfer tags (e.g., a success rate of 70%),” wherein a “success rate” encompasses one or more metric scores… associated with one or more validation metrics). However, Pedone fails to teach validation metrics comprising at least one of: precision metric, recall metric, F1-score metric, and confusion metric.
Hertz teaches metrics comprising at least one of: precision metric, recall metric, F1-score metric, and confusion metric (¶137 “Given a free text document, we first perform named entity recognition (NER) on the document to extract various types of entities… by adopting a set of in-house natural language processing techniques that include both rule-based and machine learning algorithms,” ¶141 “The algorithm is precision-oriented to avoid introducing too many false positives into the knowledge graph… relation extraction is only applied to the recognized entity pairs in each document,” ¶145 “the actual entity linking step ensures that we only generate a link when there is sufficient evidence to achieve an acceptable level of precision, i.e., the similarity between the given entity and a candidate node is above a threshold”).
Pedone and Hertz are analogous art to the claimed invention as both are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the precision metric of Hertz with the validation metric of Pedone. The motivation to do so is “to integrate… mined information from unstructured data with existing structured data and to ultimately generate insights for users based upon such integrated data” (Hertz, ¶207).
Regarding claim 9, Pedone in view of Agrawal and further in view of Hertz, and further in view of Tanniru teaches the machine-learning based (ML-based) computing method of claim 8 (and thus the rejection of claim 8 is incorporated).
Regarding the limitation further comprising adjusting, by the one or more hardware processors, the one or more hyperparameters to fine-tune the machine learning model based on one or more results of validation of the machine learning model, Pedone teaches further comprising adjusting, by the one or more hardware processors, the one or more parameters to fine-tune the machine learning model based on one or more results of validation of the machine learning model (¶232 “when the Secure Agent can detect nine (9) out of the ten (10) known transfer tags, the Secure Agent is deemed to be performing sufficiently,” ¶250 “The accuracy [of] the Secure Score is continuously refined by comparing the Secure Score to known instance of transfer tags observed… The parameters (e.g., node weights and formulas) of the Secure Agent are continuously adjusted to improve the accuracy of transfer tag detection,” ¶252 “The weighting coefficients of the neural network software applications are adjusted to reduce the error rates, and the process is run iteratively to train the neural network”). However, Pedone fails to teach adjusting… the one or more hyperparameters…
Tanniru teaches adjusting hyperparameters (Fig. 2 – 240, ¶68 “The machine learning system may tune each machine learning algorithm using one or more hyperparameter sets 240 (e.g., based on operator input that identifies hyperparameter sets 240 to be used, based on randomly generating hyperparameter values, and/or the like)”).
Pedone and Tanniru are analogous art to the claimed invention as both are in the same field of endeavor of machine learning. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the hyperparameter adjustment of Tanniru with the fine-tuning of Pedone. The motivation to do so is to “improve speed and efficiency of the [image-based document processing] and conserve computing resources (e.g., processing resources, memory resources, communication resources, and/or the like), networking resources, and/or the like” (Tanniru, ¶56).
Regarding claim 10, Pedone in view of Agrawal and further in view of Hertz, and further in view of Tanniru teaches the machine-learning based (ML-based) computing method of claim 7 (and thus the rejection of claim 7 is incorporated).
Pedone further teaches further comprising re-training, by the one or more hardware processors, the machine learning model over a plurality of time intervals based on one or more training data (Fig. 6 – 604-606, 612, ¶96 “In step 604, user evaluation data is received, collected, accessed, or otherwise acquired and entered as can be termed data ingestion… newly trained models are tied to a set of versioned data… If the preprocessing step 606 is updated with newly ingested data, an updated model will be generated,” ¶98 “A model architecture is trained in the iterative training and testing loop… Subsequent iterations of the model training, in step 612, may be conducted with updated weights in the calculations”), wherein re-training the machine learning model over the plurality of time intervals comprises: receiving, by the one or more hardware processors, the one or more training data associated with at least one of: the one or more first information extracted from the one or more electronic documents and the metadata extracted from the one or more first information (¶227 “An example system for securing and validating electronic transfer instruments and detecting errors… includes a Secure Agent software engine that processes electronic transfer instruments… to detect potential errors,” ¶230 “the Secure Agent analyzes live production data to train one or more neural networks that implement the artificial intelligence technology of the Secure Agent,” ¶245 “Image Processing Module can also detect and log irregularities in the transfer instrument image data… The Image Processing Module stores the irregularities as flagged input data within the transfer data that is passed to the Secure Agent,” ¶251 “After a transaction is processed, the results of the analysis by the Secure Agent are stored to a historical secure agent database record is created and stored… The labeling analysis determines whether transfer tags were present in a particular transaction but not caught by the Secure Agent. The historical labeling data is stored to the historical resource initialization database record to create a training data set”).
Pedone further teaches adding, by the one or more hardware processors, the one or more training data with the one or more training datasets to generate one or more updated training datasets (Fig. 6 – 604-606, 612, ¶¶96, 98, 251, wherein “newly ingested data” and data accumulated “to a historical secure agent database” implies adding… the one or more training data… to generate one or more updated training datasets).
Pedone further teaches re-training, by the one or more hardware processors, the machine learning model to correlate the one or more feature vectors associated with at least one of: the one or more first information and the metadata, with the one or more first linkage values (¶¶127, 241, 247, 251-252 all as explained above with respect to claim 7, ¶250 “The parameters (e.g., node weights and formulas) of the Secure Agent are continuously adjusted to improve the accuracy of transfer tag detection”), wherein the one or more confidence scores are generated based on re-training the machine learning model (¶250 “numerous other types of data are processed by the Secure Agent to generate a Secure Score. The accuracy [of] the Secure Score is continuously refined by comparing the Secure Score to known instance of transfer tags observed”).
Pedone further teaches executing, by the one or more hardware processors, the re-trained machine learning model in a linkage value determining subsystem to determine the one or more first linkage values for the one or more first information associated with the one or more electronic documents (¶6 “The systems can readily extract and process transfer data,” ¶7 “executable code causes the processor to transmit system configuration data to a network computer that… returns end user data… The system returns end user data, such as a user identification or user product identification (e.g., account numbers),” wherein “executable code” within a “system” that causes the system to return “end user data, such as user identification” encompasses a linkage value determining subsystem when given its broadest reasonable interpretation, ¶9 “The computer extracts transfer data elements from the transfer instrument image where the computer uses the map index to locate a component, and then reads the machine encoded content elements from the component. The computer then converts groups of machine encoded content elements to an extracted transfer data element. The computer utilizes the transfer data, end user data…to secure the transfer instrument by detecting indicators of fraud, errors, or inconsistencies called transfer tags… the computer determines a Secure Score that corresponds to a likelihood that the transfer instrument is, or is not, fraudulent or erroneous such that it should not be processed. The system includes a Secure Agent that determines the Secure Score to detect indicators of fraud or errors through the use of artificial intelligence and machine learning technology,” ¶252 “Once the neural network software applications are trained in a non-production environment, the software applications are uploaded to a production environment campaign manager to process user data in real-time,” wherein executing… the re-trained machine learning model is implicit when deploying a continuously updated model in a real-time “production environment”).
Regarding claim 17, Pedone further teaches further comprising a training subsystem configured to train the machine learning model, wherein in training the machine learning model, the training subsystem is configured to: obtain one or more labelled datasets from the one or more first databases, wherein the one or more labelled datasets comprise the one or more first information extracted from the one or more electronic documents (¶8, Fig. 1 – 124, 134, 136, 146, ¶¶37, 42, Fig. 13, ¶¶227, 251 all as explained above with respect to claim 7, ¶252 “The training data set is input to neural network software applications and machines that implement the Secure Agent. The outputs of the neural network software applications are evaluated to determine whether the outputs match the historical labeling data within pre-defined error rates. The weighting coefficients of the neural network software applications are adjusted to reduce the error rates, and the process is run iteratively to train the neural network,” wherein “neural network software applications” function in substantially the same way as a training subsystem, as described above in the interpretations under 112(f)).
Claims 17-19 are system claims that contain similar limitations to the methods of claims 7-8 and 10, respectively. Therefore, claims 17-19 are rejected under substantially the same rationale as claims 7-8 and 10, respectively.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM M LEE whose telephone number is (571)272-4761. The examiner can normally be reached Mon-Fri. 8am-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, Cesar Paula can be reached at (571)272-4128. 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.
/WILLIAM MICHAEL LEE/
Examiner, Art Unit 2145
/CESAR B PAULA/Supervisory Patent Examiner, Art Unit 2145
Prosecution Insights