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 .
Acknowledgement
Acknowledgement is made of applicant’s amendment made on 07/20/2026. Applicant’s submission filed has been entered and made of record.
Status of the Claims
Claims 1-20 are pending.
Response to Applicant’s Arguments
In response to “The amended claims clarify this architectural feature, requiring that the plurality of clients from whose past transaction data the training data is generated be "unrelated to one another." This limitation directly recites the technical mechanism by which the model achieves the coverage needed to resolve entity identity without external assistance, distinguishing the claimed training architecture from one that trains on a single organization's transaction history” and “The claims amended herein clarify this point by specifying that the transaction data payload is "submitted for authorization of a pending transaction." The transaction evaluation is also expressly required to be conducted "within a predefined response deadline of a transaction network," the precise timing constraint the specification identifies as the technical limitation the invention overcomes. The ability to skip an Internet search and additional review is not incidental. It is an architectural consequence of training the model on cross-client data, which provides the coverage needed to identify named entities without external assistance. The present claims are therefore directed to a specific improvement in a technical process, not to an abstract commercial outcome. Therefore, the present claims are not directed to an abstract idea”.
The Supreme Court and the CAFC distinguished between (1) computer-functionality improvements from the (2) uses of existing computers as tools in aid of processes focused on abstract ideas. Electric Power Grp., L.L.C. v. Alstom SA, 830 F.3d 1350, 1354 (Fed. Cir. 2016) (“…we relied on the distinction made in Alice between, on one hand, computer-functionality improvement and, on the other, uses of existing computers as tools in aid of processes focused on “abstract ideas”…”).
In the instant claims, the named-entity-identification model is effectively a computer used as a tool to identify an entity in order to conduct a transaction evaluation to determine whether to approve a pending transaction based on transaction rule.
Here, conducting a transaction is classified under certain methods of organizing activities because it corresponds to fundamental economic principles or practices for sales activities. Further, making transaction evaluations to make a determination of whether to approve a pending transaction based on transaction rule corresponds to following rules and making mental evaluation performed in human minds.
Therefore, the focus of the claims is not on improving a specifically asserted computer functionality, but rather on the use of a computer model as a tool to conduct a fundamental economic practice of making a transaction sales.
Finally, unlike the allowed claims in the parent application that specifically asserted a transaction network comprising authorizer server and rule management server performing respective specifically asserted processes that amounted to an inventive concept like implementing filtering content in a specifically asserted server client structure of BASCOM, the instant claims do not specifically assert any particular transaction network.
Therefore, the claims are patent ineligible because the claims focused on fundamental economic practice of conducting a sales transaction that do not integrate the fundamental economic practice into a practical application by specifically asserting a technological improvement or reciting an inventive concept by implementing the fundamental economic practice on a particular server client transaction network structure.
In response to “But identifying future scheduled recurring payments is not receiving a transaction data payload submitted for authorization of a pending transaction. Evans is a subscription management system for individual consumers. It analyzes a consumer's historical payment records to detect recurring subscription services that appear to be unused, and then presents cancellation, suspension, or modification recommendations to the consumer via an interactive dashboard. Evans therefore operates asynchronously and entirely outside the authorization process. There is no incoming authorization request in Evans, no transaction payload being submitted for approval, and no approval or denial decision made in response to an authorization event”.
Barbour teaches a consumer may make a purchase of a product from a merchant using a credit card (Col 11, Rows 32-35) where a point of sale system communicates the credit card information to a financial instruction associated with the merchant to authorize the purchase (Col 11, Rows 38-42; i.e., “incoming authorization request”). The merchant’s financial institution communicates with consumer’s financial institution to determine whether to authorize the transaction (Col 11, Rows 43-47; i.e., “transaction payload being submitted for approval”). The consumer’s financial institution then communicates a “yes” or “no” to the merchant’s financial institution so as to communicate either an authorization for the purchase or a denial to the point of sale system (Col 11, Rows 49-57; “approval or denial decision made in response to an authorization event”).
In a parallel process, Evans teaches analyzing electronic payment transactions from an account associated with a user to identify payments to a merchant on a recurring basis (¶59 and Fig. 4). In particular, a transaction analysis circuit 246 identifies subscription-based goods or services that the user is paying for on a recurring basis by identifying recurring payments that are scheduled to be made to automatically recur on a daily, monthly, bi-monthly, quarterly, yearly basis (¶43), in which the user had scheduled an electronic transaction with the merchant system to request a credit card payment (¶19).
Therefore, in Evans, electronic payment transactions correspond to the incoming authorization request much like the consumer’s credit card purchase of a product in Barbour, which required a parallel process of submitting transaction payload for approval and a parallel process of approval or denial decision made in response to an authorization event.
In response to “The claim language is amended herein to clarify this distinction between a future payment and a pending authorization that must happen in real-time, i.e., within a predefined response deadline of a transaction network. The amended recitation of "submitted for The claim language is amended herein to clarify this distinction between a future payment and a pending authorization that must happen in real-time, i.e., within a predefined response deadline of a transaction network. The amended recitation of "submitted for authorization of a pending transaction" directly captures the incoming authorization request context that neither Evans nor Barbour addresses. The further recitation of "within a predefined response deadline" establishes a concrete timing constraint consistent with the specification's description that transaction approval must occur within a system-imposed response deadline. See Specification at 54. Evans' asynchronous subscription monitoring system cannot satisfy that constraint”.
In view of such amendment to the independent claims, rejections under Barbour and Evans have been withdrawn. Upon further search and consideration, please see details of a new combination of references cited below.
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 USC 101 as directing toward non-statutory subject matter.
Claim 1 recites a computer implemented method (“process”). Claim 11 recites a non-transitory computer readable medium storing instructions for a processor to execute (“manufacture”). Claim 20 recites a system comprising a processor and memory (“machine”).
To distinguish ineligible claims that merely recite a judicial exception from eligible claims that require an implementation of judicial exception, the Supreme Court uses a two-step framework: Step One (Step 2A), determine whether the claims at issue are directed to one of those patent-ineligible concepts; and Step Two (Step 2B), if so, ask “what else is there in the claims?” to determine whether the additional elements transform the nature of the claim into a patent eligible application. Alice Corp. Pty. Ltd. v. CLS Bank Int’l., 134 S. Ct. 2347, 2355 (2014).
Step One (Step 2A) is a two prong test that requires the determination of whether the claims at issue are directed to an enumerated patent ineligible concept. See MPEP 2106.04.
Specifically, Step 2A Prong (1) requires the determination of the specific limitations in the claim under examination (individually or in combination) that the examiner believes recites an abstract idea and determining whether the identified limitations falls within the subject matter groupings of abstract ideas enumerated. See MPEP 2106.04(a).
The enumerated patent ineligible concepts comprising:
(a) Mathematical Concepts – mathematical relationships, mathematical formulas or equations, mathematical calculations;
(b) Certain methods of organizing human activity – fundamental economic principles / practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules / instructions) and
(c) Mental processes – concepts performed in the human mind (including an observation, evaluation, judgment, opinion). See MPEP 2106.04(a).
If the claim recites an enumerated patent ineligible concept, then Prong (2) of Step One (Step 2A) requires the determination of whether the claim integrates the patent ineligible concept into a practical application. Individually and in combination, identifying whether there are any additional elements recited in the claim beyond the judicial exceptions and evaluating those additional elements to determine whether they integrate the exception into a practical application, using one or more of the considerations laid out by the Supreme Court and the Federal Circuit. See MPEP 2106.04(d).
Under Step Two (Step 2B), if the claim does not integrate the ineligible concept into a practical application and therefore directed to a judicial exception, evaluate whether the claim provides an inventive concept by determining whether there are additional elements, individually and in ordered combination, amount to significantly more than the exception itself. See MPEP 2106.04.
Step 2A Prong (1)
The “directed to” inquiry does not ask whether the claims involve a patent ineligible concept but, considered in light of the specification, whether the claim as a whole is directed to excluded subject matter or directed to an improvement to computer functionality. Enfish L.L.C. v. Microsoft Corp., 822 F.3d 1327, 1335 (Fed. Cir. 2016).
Therefore, Prong (1) of Step 2A requires identifying specific limitations in the claims that recites (“describes” or “set forth”) an abstract idea and determine whether the identified limitations falls within the subject matter groupings of abstract ideas enumerated. See MPEP 2106.04 (“Thus, it is sufficient for this analysis for the examiner to identify that the claimed concept (the specific claim limitation(s) that the examiner believes may recite an exception) aligns with at least one judicial exception”).
In particular, MPEP 2106.04(a)(2) states “a claim that recites a mathematical calculation, when the claim is given its broadest reasonable interpretation in light of the specification, will be considered as falling within the "mathematical concepts" grouping. A mathematical calculation is a mathematical operation (such as multiplication) or an act of calculating using mathematical methods to determine a variable or number, e.g., performing an arithmetic operation such as exponentiation”.
Under Prong (1), claim 1 recites a computer-implemented method, comprising:
(1) storing past transaction data associated with a plurality of clients, the plurality of clients being unrelated to one another;
(2) generating training data from the past transaction data, the training data comprising labels of named entities associated with the past transaction data and noisy data fields in the past transaction data;
(3) training a named-entity-identification model using the labels of the named entities and the noisy data fields in the training data generated from the past transaction data;
(4) receiving a transaction rule from a particular client;
(5) receiving a transaction data payload submitted for authorization of a pending transaction associated with the particular client;
(6) identifying a noisy data field in the transaction data payload, the noisy data field including a representation of a particular named entity and one or more irrelevant strings;
(7) inputting the noisy data field into the named-entity-identification model to identify the particular named entity;
(8) applying the transaction rule to the particular identified named entity; and
(9) conducting, within a predefined response deadline of a transaction network, a transaction evaluation to determine whether to approve the pending transaction based on the transaction rule.
Claim 11 recites a non-transitory computer-readable medium configured to store computer code comprising instructions, wherein the instructions, when executed by a processor, cause the processor to implement the method of claim 1.
Claim 20 recites a system comprising: a processor; memory configured to store computer code comprising instructions, wherein the instructions, when executed by the processor, cause the processor to implement the method of claim 1.
Individually, in view of the specification US 2025/0165713 A1 at ¶77, under the broadest reasonable interpretation, steps (1)-(3) correspond to training a component trained to identify or recognize named entities based on stored past data. Therefore, steps (1)-(3) correspond to collecting data (collect past transaction data to generate training data), recognizing certain data within the collected data set (train a component to recognize or identify named entity), and storing that recognized data in a memory (store past transaction data), which are drawn to an abstract idea. TLI, 823 F.3d at 613 (citing Content Extraction and Trans. v. Wells Fargo Bank, 776 F.3d 1343, 1347 (Fed. Cir. 2014)). See also See MPEP 2106.04(a)(2)IIIA (“a claim to “collecting information, analyzing it, and displaying certain results of the collection and analysis," where the data analysis steps are recited at a high level of generality such that they could practically be performed in the human mind”).
Individually and under the broadest reasonable interpretation, steps (4)-(5) correspond to collecting information, including when limited to particular content (i.e., transaction data payload and transaction rule) that does not change its character as information, are within the realm of abstract idea. Electric Power Grp., L.L.C. v. Alstom SA, 830 F.3d 1350, 1353 (Fed. Cir. 2016).
Individually and under the broadest reasonable interpretation, steps (6)-(7) correspond to recognizing certain data within the collected data set (i.e., recognize noisy data field and recognize named entity). Analyzing information by steps people go through their minds, without more, are essentially mental processes within the abstract-idea category. Electric Power Grp., 830 F. 3d at 1354.
Individually and under the broadest reasonable interpretation, steps (8)-(9) correspond to conducting a transaction evaluation to determine whether to approve the pending transaction based on collected information corresponding to transaction rule and recognized named entity. Analyzing information by steps people go through their minds, without more, are essentially mental processes within the abstract-idea category. Id.
As an ordered combination, steps (1)-(8) described a series of mental analysis of using transaction rule ad named entity identification rule on named entities / merchants to determine whether to approve a pending transaction or business relation with the named entities / merchants.
Thus, Claims 1, 11, and 20 described patent ineligible subject matter enumerated under category (b) Certain methods of organizing human activity – fundamental economic principles / practices (including hedging, insurance, mitigating risk); commercial or legal interactions (including agreements in the form of contracts; legal obligations; advertising, marketing or sales activities or behaviors; business relations); managing personal behavior or relationships or interactions between people (including social activities, teaching, and following rules / instructions) and (c) Mental processes – concepts performed in the human mind (including an observation, evaluation, judgment, opinion)..
Step 2A Prong (2).
Under Prong (2) of Step 2A, the goal is to determine whether the claim is directed to the recited exception by evaluating whether the claim as a whole integrates the recited judicial exception into a practical application of the exception. See MPEP 2106.04II(A).
In particular, evaluating integration into a practical application requires identifying whether there are any additional elements recited in the claim beyond the judicial exception and evaluating those additional elements, individually and in combination, to determine whether they integrate the exception into a practical application, using one or more of the considerations laid out by the Supreme Court and the Federal Circuit (“CAFC”). See MPEP 2106.04(d).
The Supreme Court held that when a claim containing a mathematical formula (i.e., an abstract idea) implements or applies that math formula / abstract idea in a structure or process which, when considered as a whole, is performing a function which the patent laws were designed to protect (e. g., transforming or reducing an article to a different state or thing), then the claim satisfies the requirements of §101. Diamond v. Diehr, 450 U.S. 175, 192 (1981); See MPEP 2106.04(d)I (“Implementing a judicial exception with, or using a judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim, as discussed in MPEP 2106.05(b)”). See also Gottschalk v. Benson, 409 U.S. 63, 70 (1972) (“Transformation and reduction of an article "to a different state or thing" is the clue to the patentability of a process claim that does not include particular machines”).
In one example, the CAFC applied Alice inquiry to ask whether the focus of the claims is on the specific asserted improvement in computer capabilities (i.e., the self-referential table for a computer database) or instead, on a process that qualifies as an abstract idea for which computers are invoked merely as a tool. Enfish L.L.C. v. Microsoft Corp., 822 F.3d 1327, 1335-36 (Fed. Cir. 2016).
In Enfish, the claims were specifically directed to a self-referential table for a computer database. Id. at 1337. In particular, the claim language required a four step algorithm specifically directed to a self-referential table for a computer database that improves upon prior art information search and retrieval systems by employing a flexible, self-referential table to store data. Id. at 1336-37. CAFC determined that the plain focus of the claims was on an improvement to computer functionality itself (i.e., the self-referential table for a computer database), not on economic or other tasks for which a computer is used in its ordinary capacity. Id at 1335-36.
Therefore, the focus of the claims is on a specific asserted improvement in computer capabilities (i.e., the self-referential table for a computer database), not on economic or other tasks for which a computer is used in its ordinary capacity. Id. at 1336. See also MPEP 2106.04(d)I (“an improvement in the functioning of a computer or an improvement to other technology or technical field, as discussed in MPEP 2106.04(d)(1) and 2106.05(a)”).
In another example, in McRO, the CAFC noted that prior art method of generating morph weight set with values between “0” and “1” for computer animation of facial expressions are manually determined. McRO, 837 F.3d at 1304-5. The claimed improvement in McRO allows computers to produce “accurate and realistic lip synchronization and facial expressions in animated characters” that previously could only be produced by human animators through the automated use of rules, rather than artists, to set the morph weights and transitions between phonemes. Id. at 1313.
Specifically, the claims are directed to the incorporation of claimed rules, not the use of the computer that improved existing technological process by allowing automation of further tasks that goes beyond merely organizing existing information into a new form. Id. at 1314-15.
In other words, the claimed process uses a combined order of specific rules that renders information into a specific format that is then used and applied to create a sequence of synchronized, animated characters that prevent pre-emption of all processes for achieving automated lip-synchronization of 3-D characters. Id. at 1315. Therefore, the CAFC held that the ordered combination of claimed steps, using unconventional rules that relate sub-sequences of phonemes, timing, and morph weight sets is patent eligible. Id. at 1302-3.
The key here, as noted by the CAFC, is that the Supreme Court in Diehr looked to how the claims "used that equation in a process designed to solve a technological problem in `conventional industry practice.'" Id. at 1312. When looked at as a whole, "the claims in Diehr were patent eligible because they improved an existing technological process, not because they were implemented on a computer." Id. at 1312-13.
On the other hand, in Intellectual Ventures I, the CAFC held that the claims were directed to an abstract idea of tracking financial transactions to determine whether they exceed a pre-set spending limit (i.e., budget). Intellectual Ventures I L.L.C. v. Capital One Bank, 792 F.3d 1363, 1367 (Fed. Cir. 2015). The CAFC determined that while the claims recited budgeting using a communication medium (broadly including the internet and telephone networks), a database, and a user profile, these limitations did not render the claims any less abstract because they are all generic computer elements that amounted to instructing one to “apply” an abstract idea and reciting no more than generic computer elements performing generic computer tasks. Id. At 1367-68.
In the instant application, Claims 1, 11, and 20 recited steps (1)-(9) described a series of steps to use transaction rules and named entity identification rules to organize past transaction data of clients into patterns to identify one or more named entities / merchants in the list selected by the client and to analyze transaction data payload to determine whether the particular named entity / merchant is allowed to do business with the client. In other words, steps (1)-(9) manipulate or organize existing / past transaction data into a new form or patterns in order to determine whether named entities / merchants in the pending transaction data payload associated with the client may do business with the client.
In contrast to the four step algorithm that described a specifically asserted self-referential table that improved database search and retrieval functions in Enfish, the claims do not describe limitations that optimize any specifically asserted computer technology.
Further, unlike the incorporation of claimed rules to automate the determination of morph weights that improved existing technological process by allowing automation of further tasks that goes beyond merely organizing existing information into a new form in McRO, organizing existing / past transaction data into a new form or patterns to conduct business transaction between named entities / merchants and client do not improve or automate any specifically asserted technological process.
Rather, much like the abstract idea of tracking financial transactions to determine whether they exceed a pre-set spending limit along with recitation of a database, user profile, and a communication medium in Intellectual Ventures I, requiring a computer to store and collect past transaction data, to train / generate a computer component to recognize or identify named entities from pending transaction data payload which are essentially a mental process, to thereafter determine whether to approve a pending transaction or business relation with the named entities / merchants recite no more than generic computer elements and “apply” the abstract idea, however developed.
Therefore, claims 1, 11, and 20 are directed to the abstract idea / mental process of using transaction rule ad named entity identification rule on named entities / merchants to determine whether to approve a pending transaction or business relation with the named entities / merchants.
Step 2B Inventive Concept.
The Guideline stated that if the additional elements do not integrate the exception into a practical application, then the claim is directed to the recited judicial exception, and requires further analysis under Step 2B where it may still be eligible if it amounts to an “inventive concept”. See MPEP 2106.04IIA and MPEP 2106.05.
Further, an inventive concept can be found in the non-conventional and non-generic arrangement of known conventional pieces. BASCOM Global Internet Servs. v. AT&T Mobility, 827, F3d 1341, 1350 (Fed. Cir. 2016).
In BASCOM, the CAFC held that filtering content is an abstract idea because it is a longstanding, well-known method of organizing human behavior similar to concepts previously found to be abstract. BASCOM, 827 F.3d at 1348. However, the CAFC determined that the claims did not merely recite filtering content along with the requirement to perform it on the internet or on a set of generic computer components, nor did the claims preempt all ways of filtering content on the internet. Id. at 1350.
Rather, the inventive concept described and claimed was the installation of a filtering tool at a specific location, remote from the end-users, with customizable filtering features specific to each end user that gives the filtering tool both the benefits of a filter on a local computer and the benefits of a filter on an internet service provider “ISP” server. Id. By taking a prior art filter solution (one size fits all filter at internet service provider “ISP” server) and making it more dynamic and efficient (providing individualized filtering at the ISP server), the claimed invention improves the performance of the computer system itself. Id. at 1351.
On the other hand, implementation via computers does not offer a meaningful limitation beyond generally linking the use of an abstract idea to a particular technological environment. Alice, 134 S. Ct. at 2360 (“Nearly every computer will include a “communications controller” and “data storage unit” capable of performing the basic calculation, storage, and transmission functions required by the method claims”). Intellectual Ventures I L.L.C. v. Capital One Bank, 792 F.3d 1363, 1370-71 (Fed. Cir. 2015) (“Steps that do nothing more than spell out what it means to “apply it on a computer” cannot confer patent-eligibility).
Similarly, limiting an abstract idea to one field of use do not convert otherwise ineligible concept into an inventive concept. Intellectual Ventures I L.L.C. v. Erie Indem. Co., 850 F.3d 1315, 1328 (Fed. Cir. 2017). Neither does adding computer functionality to increase the speed or efficiency of the process confer patent eligibility on an otherwise abstract idea. Intellectual Ventures I, 792 F.3d at 1367 (citing Bancorp Servs., LLC v. Sun Life Insurance Co. of Can., 687 F.3d 1266, 1278 (Fed. Cir. 2012) (“The fact that the required calculations could be performed more efficiently via a computer does not materially alter the patent eligibility of the claimed subject matter”)).
Individually, in the instant application, claim 1 recites a computer. Claim 11 recites non-transitory computer readable medium storing computer code instructions for execution by a processor to execute the method of claim 1. Claim 20 recites a system comprising processor and memory to implement the method of claim 1.
Such individual recitation of generic computer components (processor, computer program, storage) are purely functional and generic because nearly every computer will include such processor and data storage unit capable of performing basic calculation necessary for data calculation / analysis, storage, and transmission.
As an ordered combination, unlike BASCOM that describes an unconventional combination of a conventional ISP server with a customized filter specific to each user that is remote from end-users to provide both the benefits of a filter on a conventional local computer and the benefits of a filter on the conventional ISP server, implementing steps (1) –(9) using transaction rule ad named entity identification rule on named entities / merchants to determine whether to approve a pending transaction or business relation with the named entities / merchants do not involve a unconventional combination of conventional pieces because the combination amounts to “apply it on a computer” or limiting the application of (1)-(9) to the field of computers.
To the extent that implementing steps (1)-(9) on a processor / computer in the field of computers results in reduction in memory requirement and computational requirement, merely adding computer functionality to increase the speed or efficiency of approving a pending business transaction between user and merchant does not confer patent eligibility on an otherwise abstract idea.
Dependent claims also failed to integrated claims 1 and 11 into a practical application under step 2A or an inventive step under step 2B for the following reasons:
Claims 2-3 and 12-13 are directed toward collecting data, analyzing data, and recognizing certain data within the collected data set. As previously noted, recognizing certain data within the collected data set, and storing that recognized data in a memory are drawn to an abstract idea. TLI, 823 F.3d at 613. See also See MPEP 2106.04(a)(2)IIIA (“a claim to “collecting information, analyzing it, and displaying certain results of the collection and analysis," where the data analysis steps are recited at a high level of generality such that they could practically be performed in the human mind”).
Claims 4 and 14 are directed toward mathematical analysis of labeled training samples to supervise the training of a math model, optimize the mathematical model using mathematical method of gradient descent, and using mathematical cross-entropy loss function to evaluate the accuracy of the mathematical model. Analyzing information by mathematical algorithms, without more, are essentially mental processes within the abstract-idea category. Electric Power Grp., 830 F.3d at 1354.
Claims 5 and 15 are directed toward collecting data, limited to particular content including client specific requirements. Collecting data, recognizing certain data within the collected data set, and storing that recognized data (client specific rules) in a memory are drawn to an abstract idea. Content Extraction, 776 F.3d at 1347.
Similarly, claims 6 and 16 are directed toward recognizing data.
Claims 7 and 17 are directed to calculating multi-dimensional vector representation of collected data. Analyzing information by mathematical algorithms, without more, are essentially mental processes within the abstract-idea category. Electric Power Grp., 830 F.3d at 1354.
Claims 8 and 18 describe a particular neural network structure. However, as detailed below, this neural network structure is well known and conventional.
Further, in USPTO’s Memo on 2024 Updated Guidance on AI and Subject Matter Eligibility issued July 16, 2024 describing example 48 on pp. 14-15 regarding a method to separate speech signals from different sources to recognize human speech command from background noise by using a deep neural network (DNN) to promote separation of the features during clustering. See p. 15, ¶2. Specifically, the DNN learns high level feature representations of the signal x by mapping the feature representations to the embedding space comprising the DNN converting feature representations Xt, obtained from spectrograms St and corresponding feature matrices FMt, into multi-dimensional embedding vectors V and assigning the embedding vectors V to TF bins as a global function of the input signal (V= fθ(X), where fθ represents a function of the DNN). See p. 15, ¶5.
According the Memo under Step 2A, Prong One, a claim comprising a step of using a deep neural network (DNN) to determine embedding vectors V using the formula V= fθ(X), where fθ represents a function of the DNN describes a mathematical calculation and therefore the claim “set forth” or “describes” a judicial exception. p. 19, ¶5.
Under Step 2A, Prong Two, since there is no detail about a particular DNN or how the DNN operates to derive the embedding vectors other than that it is being used to determine the embedding vectors, the DNN is used to generally apply the abstract idea of performing mathematical calculation using recited mathematical equation without placing any limitation on how the DNN operates to derive the embedding vectors as a function of the input signal. p. 20, ¶2.
In particular, the disclosure identifies a technical problem encountered in the field of speech separation and provides an improvement over existing speech separation methods by determining embedding vectors as a function of the input signal, partitioning those vectors into clusters, and synthesizing a reconstructed mixed speech signal based on these clusters. p. 20, ¶3. The claim, however, only requires determining the embedding vectors and therefore does not reflect the improvement discussed in the disclosure. Id.
Similarly, the neural network structure of claims 8 and 18 comprising convolution layer and the recurrent layer functioned to extract mathematical vectors representing feature and temporal pattern with fully connected layers functioned to perform classification. Much like the DNN in the example 48 above, the neural network architecture functioned pure to perform the abstract process of data recognition.
Put simply, while there is a particular neural network structure, the neural network structure functioned to perform abstract process of data recognition as part of organizing human activity to conduct a business transaction that do not amount to a specifically asserted technological process or a specifically asserted improvement corresponding to a novel neural network structure (see prior art rejection of claims 8 and 18).
Claims 9 and 19 perform additional data recognition process to make an evaluation of whether the business transaction violate a predefined threshold. Such process still corresponds to conducting business or organizing human activity.
Claim 10 corresponds to collecting additional data limited to particular content of new training data and patterns of data recognition (i.e., merchant recognition) but does not change its character as data or information.
Claim 20 describe pattern analysis of past transaction data to determine a named entity identification rule in order to conduct an Internet search to determine an identity of a candidate named entity. Much like the recitation of budgeting using a communication medium (broadly including the internet and telephone networks) in Intellectual Ventures I, performing pattern analysis to perform an internet search amounted to apply the abstract idea along with generic computer components.
In conclusion, the process of claims 1-20 focused on an abstract idea pertaining to certain methods of organizing human activity or organizing business relation by using transaction rule and named entity identification rule on named entities / merchants to determine whether to approve a pending transaction or business relation with the named entities / merchants.
Therefore, Claims 1-20 are not eligible for a patent.
Nonstatutory Double Patent Rejections
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the "right to exclude" granted by a patent and to prevent possible harassment by multiple assignees. See In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); and In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) may be used to overcome an actual or provisional rejection based on a nonstatutory double patenting ground provided the conflicting application or patent is shown to be commonly owned with this application. See 37 CFR 1.130(b).
Effective January 1, 1994, a registered attorney or agent of record may sign a terminal disclaimer. A terminal disclaimer signed by the assignee must fully comply with 37 CFR 3.73(b).
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/forms/. The filing date of the application will determine what form should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/process/file/efs/guidance/eTD-info-I.jsp.
Claims 1, 5, 9, 11, 15, and 19-20 are rejected under the judicially created doctrine of obviousness-type double patenting as being unpatentable over the claims of US 12217004 B2 in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1).
The respective limitations of claims 1+6, 7+17, and 19 of US 12217004 B2 fully encompass the limitations of claims 1, 11, and 20 in the instant application except for “the plurality of clients being unrelated to one another” and “conducting the transaction evaluation within a predefined response deadline of a transaction network”.
Jamkhedkar discloses conducting, within a predefined response deadline of a transaction network, a transaction evaluation to determine whether to approve the pending transaction based on a transaction rule (¶69, perform verification to communicate approval or denial to the merchant within merchant’s timeout of 2.5 seconds).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to conducting the transaction evaluation within a predefined response deadline of a transaction network in order to provide a prediction of approval or denial (Barbour, Col 11, Rows 53-54) in a timely manner in accordance with a desired quality of service parameter (Jamkhedkar, ¶33).
Evans discloses a system (Figs. 1-2) comprising store past transaction data associated with a plurality of clients being unrelated to one another (¶35, account database 240 at system 104 stores transaction history data of transactions made by user 101 and by other customers).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to train the named entity identification model based on training data generated from past transaction data associated with clients unrelated to one another in order to implement machine learning analysis of transaction history of customers transacted with merchants / named entities (Evans, ¶36).
Regarding Claims 5 and 15, the claim limitations of US 12217004 B2 do not disclose wherein the instructions to receive the transaction rule from the particular client comprises instructions to:
present a graphical user interface for client input; allow the particular client to select from preconfigured rule templates; and dynamically update the rule based on client-specific requirements.
Evans discloses wherein the instructions to receive the transaction rule from the particular client comprises instructions to:
present a graphical user interface for client input (¶17 and ¶28 and Fig 2A, user device 102 with electronic user interface for interactive display supported by output circuit 212 and to receive user input from the user via input circuit 210);
allow the particular client to select from preconfigured rule templates (¶38 provide predefined / automatically generated rules to the user to include user defined exceptions in which the user defines situations in which a rule may be ignored); and
dynamically update the rule based on client-specific requirements (¶39, upon receiving user defined automatically generated rules, index the rules and store the rules in the rules database 252).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to present a graphical user interface for client input, allow the particular client to select from preconfigured rule templates, and dynamically update the rule based on client-specific requirements to allow users to defined exceptions (Evans, ¶38).
Regarding Claims 9 and 19, the claim limitations of US 12217004 B2 do not disclose wherein the instructions to apply the transaction rule comprises instructions to: cross-reference the identified named entity with a list stored in a database and trigger an alert responsive to the transaction violates a predefined threshold.
Evans discloses wherein the instructions to apply the transaction rule comprises instructions to:
cross-reference the identified named entity with a list stored in a database and trigger an alert responsive to the transaction violates a predefined threshold (¶37, alert user of a recurring payment for a good or service (i.e., pending transaction involving an entity) that violate one or more of the rules; ¶38, for example, after cross referencing a current merchant of the pending transaction to user defined exceptions to a general rule (i.e. whitelists) for particular merchants, execute the general rule to notify the user when an amount of time of disuse of the good or service provided by the current merchant exceeds a particular threshold value because the current merchant is not one of the particular merchants on the whitelists).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to apply the transaction rule comprises instructions to: cross-reference the identified named entity with a list stored in a database and trigger an alert responsive to the transaction violates a predefined threshold in order to notify a user of a pending transaction violates one or more user defined rules or exception to said rules (Evans, ¶37).
Claims 2, 7, 10, 12, and 17 are rejected under the judicially created doctrine of obviousness-type double patenting as being unpatentable over the claims of US 12217004 B2 in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in further view of Barbour (US 11640417 B2 A1).
Regarding Claims 2 and 12, the claim limitations of US 12217004 B2 do not disclose store the past transaction data associated with the plurality of clients comprises instructions to:
retrieve transaction data payloads from a transaction terminal; filter the transaction data payloads to exclude irrelevant information; and categorize the transaction data payloads by client and merchant identifier.
Barbour discloses wherein the instructions to store past transaction data associated with a plurality of clients (Col 16, Rows 19-20) comprises instructions to:
retrieve transaction data payloads from a transaction terminal (Fig. 1B and see Col 12, Rows 5-8, information retrieval system 140 receive transaction information from cardholder’s financial institution 130);
filter the transaction data payloads to exclude irrelevant information (Col 17, Rows 10-24, parse #IPX624THUDERCLDSUB . . . Austin, Tx (07/12) to determine merchant portion of this record “THUNDERCLDSUB Austin, Tx”); and
categorize the transaction data payloads by client and merchant identifier (Col 10, Rows 1-5, associate merchant record with the transaction to provide more accurate information related to the merchant to supplement and correct information in the transaction record; Col 13, Row 61 – Col 14, Row 3, by associating existing accurate merchant listing with transaction, user / cardholder financial institution can access additional information associated with the merchant to more accurately and specifically perform tasks such as identifying and categorizing transactions associated with the merchant; per Col 11, Rows 38-42, the transaction includes consumer credit card information; i.e., historical electronic transaction records comprises transaction information such as client / consumer’s credit card information and merchant portion).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to retrieve transaction data payloads from a transaction terminal; filter the transaction data payloads to exclude irrelevant information; and categorize the transaction data payloads by client and merchant identifier in order to generate a set of historical electronic transaction records to train a neural language model (Barbour, Col 16, Rows 15-20).
Regarding Claims 7 and 17, the claim limitations of US 12217004 B2 do not disclose wherein the instructions to input the noisy data field into the named-entity-identification model comprises instructions to:
tokenize the noisy data field into substrings; embed substrings into a vector in a multi-dimensional vector space; and normalize the vector.
Barbour discloses wherein the instructions to input the noisy data field into the named-entity-identification model comprises instructions to:
tokenize the noisy data field into substrings (Col 9, Rows 35-40, determine merchant portion of the transaction record, tokenize the merchant portion to determine terms to use as a query; see e.g., Col 17, Rows 10-33, parse character strings of electronic transaction record to determine merchant portion of received transaction record as search terms);
embed substrings into a vector in a multi-dimensional vector space (Col 17, Rows 34-53, obtain an embedding / single vector for the search terms from neural language modeler 267; per Col 22, Rows 26-27, neural language model maps terms into a multi-dimensional vector space); and
normalize the vector (Col 17, Row 53 – Col 18, Row 11, generate the single vector representing the search terms by weighting each of the search terms with inverse document frequency associated with respective terms).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to tokenize the noisy data field into substrings; embed substrings into a vector in a multi-dimensional vector space; and normalize the vector to generate a compact mathematical representation capturing the semantic or syntactic characteristics of the text (Barbour, Col 10, Rows 9-15).
Regarding Claim 10, the claim limitations of US 12217004 B2 do not disclose wherein the named-entity-identification model is periodically updated to include: new training data generated from cleared transactions and enhanced features derived from recent merchant identification patterns.
Barbour discloses wherein the named-entity-identification model is periodically updated to include:
new training data generated from cleared transactions (Col 16, Rows 19-27, update or additionally train a neural language model based on specific corpus comprising set of historical electronic transaction records; per Col 16, Rows 65-66, electronic transaction records associated with completed transaction); and enhanced features derived from recent merchant identification patterns (Col 13, Rows 50-61 and Col 25, Rows 6-22, electronic transaction records corresponding to “Thundercld Sub” corrected to the more accurate and specific “Thundercloud Subs”).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to periodically update the named-entity-identification model to include new training data generated from cleared transactions and enhanced features derived from recent merchant identification patterns in order to generate specific corpus to additionally train the neural language model / named entity identification model as desired (Barbour, Col 16, Rows 23-27).
Claims 3 and 13 are rejected under the judicially created doctrine of obviousness-type double patenting as being unpatentable over the claims of US 12217004 B2 in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in further view of Relangi et al. (US 2020/0372219 A1).
The limitations of claims 1+6 and 7+17 of US 12217004 B2 encompass the limitations of identify patterns in noisy data fields based on predefined rules and assign labels to named entities using historical data.
The claim limitations of US 12217004 B2 do not disclose validate the labeled data through a feedback loop.
Relangi teaches validating labeled data through a feedback loop (Abstract, Fig. 1, steps 130 -> 140 -> “No” -> 145; ¶37 and ¶43).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to valid labeled data through a feedback loop in order to receive corrected labelled named entities from a reviewer (Relangi, ¶10).
Claims 4 and 14 are rejected under the judicially created doctrine of obviousness-type double patenting as being unpatentable over the claims of US 12217004 B2 in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in further view of Zhong et al. (US 2021/0303638 A1) and Li et al. (US 2022/0083744 A1).
The limitations of claims 6 and 17+18 of US 12217004 B2 encompass the limitations of utilize machine learning techniques with labeled training samples and iteratively optimize the model.
The claim limitations of US 12217004 B2 do not disclose utilize supervised learning techniques with labeled training samples, iteratively optimize the model using gradient descent; and evaluate model accuracy based on cross-entropy loss.
Zhong teaches training an embedding model by utilizing supervised learning techniques (¶2, using artificial neural networks, support vector machines, decision tress, naïve Bayes classifiers to analyze data sets1) with labeled training samples (¶58, train embedding model 208 based on input string-entity pairs and labels 214); optimize the model using gradient descent and evaluate model accuracy based on cross-entropy loss (¶58, use gradient descent and backpropagation and a cross entropy loss function to update parameter values of embedding model 208 to reduce error between output of the embedding model 208 and corresponding labels 214).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to supervised learning techniques with labeled training samples, optimize the model using gradient descent, and evaluate model accuracy based on cross-entropy loss in order to reduce error between output of the neural language model / embedding model and corresponding labels / categorizations in the training corpus (Zhong, ¶58).
The combination does not teach that the optimization of the model using gradient descent is iterative.
Li teaches iterative optimization through gradient descent training of a neural language model (¶24).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to perform iterative optimization of the model using gradient descent in order to iterate until there is no further improvement in the loss function (Li, ¶24).
Claims 6 and 16 are rejected under the judicially created doctrine of obviousness-type double patenting as being unpatentable over the claims of US 12217004 B2 in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in further view of Yang et al. (US 2012/0278705 A1).
The claim limitations of US 12217004 B2 do not disclose wherein the instruction to identify the noisy data field comprises instructions to apply heuristic algorithms to detect irrelevant substrings and standardize a format of extracted data fields.
Yang discloses automatically extracting meta data from unstructured electronic documents (Abstract) by applying heuristic algorithms to detect irrelevant substrings corresponding to noisy data fields (¶23, noise text line may be a text line mistaken as text line associated with header of an unstructured electronic document; ¶57, using heuristic approach to identify noise text line) and standardize a format of extracted data fields (¶57, classify text lines into different categories such as names (i.e., named entities), affiliations etc.).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to applying heuristic algorithms to detect irrelevant substrings corresponding to noisy data fields and standardize a format of extracted data fields in order to convert unstructured text into plain text (Yang, ¶57).
Claims 8 and 18 are rejected under the judicially created doctrine of obviousness-type double patenting as being unpatentable over the claims of US 12217004 B2 in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in further view of Wu et al. (US 2020/0082928 A1).
The claim limitations of US 12217004 B2 do not disclose wherein the named-entity-identification model utilizes a neural network architecture comprising: convolutional layers for feature extraction; recurrent layers for temporal pattern recognition; and fully connected layers for classification.
Wu discloses that it is well known to use a neural network architecture comprising convolutional layers for feature extraction (¶135, CNN encoding layer 1246 to generate vector representation of image input such as image messages to generate corresponding text per ¶59), recurrent layers for temporal pattern recognition (¶¶126-27, RNN-GRU encoding layer 1244 to project temporal voice inputs into vector representation), and fully connected layers for classification (¶148) of user inputs into vector space (¶121).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to implement a neural network architecture comprising convolutional layers for feature extraction; recurrent layers for temporal pattern recognition; and fully connected layers for classification of a message input comprising text, voice, and image into vector space (Wu, Fig. 12 and see ¶148). For example, such neural language model can be applied to transaction comprising text, voice, and image to generate respective multi-dimensional vectors.
Claim Rejections - 35 USC § 103
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 103 that form the basis for the rejections under this section made 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, 5, 7-12, 15, 17, and 19-20 are rejected under 35 USC 103 for being unpatentable over Barbour (US 11640417 B2 A1) in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1).
Regarding Claims 1, 11, and 20, Barbour discloses a system (Fig. 1B) comprising:
a processor (Fig. 1B, CPU);
memory configured to store computer code comprising instructions, wherein the instructions, when executed by the processor (Col 14, Rows 18-36, CPU 212 connected to data store 218 / RAM storing computer executable instructions 219 to provide search engine 251 and neural language modeler 261), cause the processor to:
store past transaction data associated with a plurality of clients (Col 14, Rows 43-47, repository 205 stores records 207 that form a corpus that may be searched through information retrieval system 201; e.g., Col 15, Rows 35-38, records 207 in the corpus comprising one record per merchant; Col 15, Rows 66-67, merchant records based on merchant portion of electronic transaction records; and Col 16, Rows 17-27, the corpus includes a specific training corpus 269 comprising a set of historical electronic transaction records);
generate training data from the past transaction data, the training data comprising labels of named entities associated with the past transaction data and noisy data fields in the past transaction data (Col 16, Rows 19-20, use a set of historical electronic transaction records as the specific training corpus 269; in view of Col 12, Row 59 – Col 13, Row 13, transaction record (e.g., Col 13, Row 13) comprising text strings of merchant entity and noise);
train a named-entity-identification model using the labels of the named entities and the noisy data fields in the training data (Col 15, Rows 46-49, trainer 265 trains neural language model 267 based on training data 269 stored in repository 205) generated from the past transaction data (Col 16, Rows 9-20, train the neural language model 267 on a specific corpus comprising a set of historical electronic transaction records);
receive a transaction data payload submitted for authorization of a pending transaction associated with the particular client (Col 11, Rows 33-43, a consumer makes a purchase of a product from a merchant using a credit card);
identify a noisy data field in the transaction data payload, the noisy data field including a representation of a particular named entity and one or more irrelevant strings (Col 13, Rows 4-7, the consumer’s financial institution wishes to identify a merchant associated with a transaction; Col 9, Rows 24-25 and Rows 35-40, merchant names in electronic transaction record is especially noisy; e.g., Col 17, Rows 10-24, receive “#IPX624THUDERCLDSUB . . . Austin, Tx (07/12)” as part of the electronic transaction record where a query term processor 255 parses the transaction record to determine merchant portion of the transaction record comprising entity “THUNDERCLD SUB”, which is noisy and inaccurate as well as irrelevant strings such as #IPX624 and (07/12));
input the noisy data field into the named-entity-identification model to identify the particular named entity (Col 6, Rows 35-50, using neural language model to find records that are associated with noisy information and usefully applied to search based on any type of noisy, and potentially inaccurate, descriptive data; Col 18, Rows 20-37, search engine 251 can embed determined search terms using vector builder 257 and neural language modeler 21 / neural language model 267 to create single mathematical representation of the search terms, measure the similarity between embedding representation of the search terms and embedding representation of merchant data record 207 to return merchant information data record 207 nearest to the search terms; e.g., Col 25, Rows 6-22, rather than the noisy “Thundercld Sub”, correctly identify the merchant name as “Thundercloud Subs”).
Barbour does not teach applying a transaction rule to a particular identified named entity and conducting, within a predefined response deadline of a transaction network, a transaction evaluation to determine whether to approve the pending transaction based on the transaction rule.
Jamkhedkar discloses receiving transaction data payload submitted for authorization of a pending transaction (¶33, receive a transaction request), receiving a transaction rule from a particular merchant / particular identified named entity (¶49 and ¶69, access merchant timeout information to run merchant specific rules), applying the transaction rule to the particular identified named entity / merchant (¶49, set timeout counter to a quality of service agreement value specific to the particular merchants), and conducting, within a predefined response deadline of a transaction network, a transaction evaluation to determine whether to approve the pending transaction based on the transaction rule (¶69, perform verification to communicate approval or denial to the merchant within merchant’s timeout of 2.5 seconds).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to identify a merchant named entity (Barbour, Col 12, Rows 4-6, consumer’s financial institution wishes to identify a merchant associated with a transaction) prior to communicating a “yes” or “no” to the merchant’s financial institution (Barbour, Col 11, Rows 38-55, merchant financial institution communicates credit card information to consumer financial institution, which communicates a “yes” or “no” to the merchant’s financial institution; compare Jamkhedkar, ¶69, verification result comprising “approve” or “deny”) by accessing a transaction rule from the merchant to conduct the transaction evaluation to determine whether to approve the pending transaction based on the transaction rule specifying a predefined response deadline of a transaction network (Jamkhedkar, ¶69, merchant specified timeout of 2.5 seconds) in order to provide a prediction of approval or denial (Barbour, Col 11, Rows 53-54) in a timely manner in accordance with a desired quality of service parameter (Jamkhedkar, ¶33).
Barbour as modified by Jamkhedkar do not disclose that past transaction data comprises a plurality of clients being unrelated to one another and receiving a transaction rule from a particular client.
Evans discloses a system (Figs. 1-2) comprising store past transaction data associated with a plurality of clients being unrelated to one another (¶35, account database 240 at system 104 stores transaction history data of transactions made by user 101 and by other customers), receiving a transaction rule from a particular client (¶17 and ¶28 and Fig 2A, user device 102 with electronic user interface for interactive display supported by output circuit 212 and to receive user input from the user via input circuit 210; ¶37, receiving user defined rules defining how to manage subscriptions for goods or services (i.e., merchants) based on user preferences; ¶38, the rules includes whitelists / exceptions to one or more of the rules or exception to the rule for a particular merchant), receiving a transaction data payload related to a pending transaction associated with the particular client (¶37 and ¶43, identifying future payments that are scheduled to automatically recur; e.g., ¶59 and Fig. 4, step 405, analyze electronic payment transactions from user’s account to identify payments to a merchant on a recurring basis), identify a particular named entity within the transaction data payload (¶44, determine whether each of the subscription based or membership based services identified by the transaction analysis circuit 246 is being used such as analyzing merchant data to determine if a corresponding goods / services provided by a merchant is being used), apply the transaction rule to the particular identified named entity and conduct a transaction evaluation to determine whether to approve the pending transaction based on the transaction rule (¶48, ¶50, and ¶62, using rules stored in rules database 252 (i.e., rules defined per ¶¶37-38 and per user preferences) to automatically cancel or suspend or modify subscription to goods / services provided by the merchant).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to identify particular named entity from noisy data field in transaction data payload related to a pending transaction using the named-entity identified model, apply transaction rule to the particular identified named entity, and conduct a transaction evaluation to determine whether to approve the pending transaction based on the transaction rule in order to identify named entity / merchant portion (Barbour, Col 9, Rows 33-35) from noisy electronic transaction records (Barbour, Col 9, Rows 19-21) and correctly apply client transaction rules to the right named entity / merchant (Evans, ¶38; compare Barbour, Col 13, Row 61 – Col 14, Row 3, by associating existing accurate merchant listing with transaction, user / cardholder financial institution can access additional information associated with the merchant to more accurately and specifically perform tasks such as identifying and categorizing transactions associated with the merchant).
Further, the predictable result of modifying Barbour’s established training of neural language model / named entity identification model using past transaction data (Barbour, Col 15, Rows 17-21) to include past transaction data associated with a plurality of clients being unrelated to one another (Evans, ¶35) would better capture semantic and syntactic characterization of noisy terms (Barbour, Col 9, Rows 24-25, merchant names are especially noisy) in electronic transaction records (Barbour, Col 16, Rows 1-8) by learning past transaction data from other clients unrelated to one another.
Further regarding claim 11, Barbour discloses a non-transitory computer-readable medium configured to store computer code comprising instructions, wherein the instructions, when executed by a processor, cause the processor to implement the method of claim 1 and functions of claim 20 (Col 14, Rows 24-33, data store 218 comprising non-transitory storage medium such as RAM to store instructions 219 to provide search engine 251 and neural language modeler 261).
Regarding Claims 2 and 12, Barbour discloses wherein the instructions to store the past transaction data associated with the plurality of clients (Col 16, Rows 19-20) comprises instructions to:
retrieve transaction data payloads from a transaction terminal (Fig. 1B and see Col 12, Rows 5-8, information retrieval system 140 receive transaction information from cardholder’s financial institution 130);
filter the transaction data payloads to exclude irrelevant information (Col 17, Rows 10-24, parse #IPX624THUDERCLDSUB . . . Austin, Tx (07/12) to determine merchant portion of this record “THUNDERCLDSUB Austin, Tx”); and
categorize the transaction data payloads by client and merchant identifier (Col 10, Rows 1-5, associate merchant record with the transaction to provide more accurate information related to the merchant to supplement and correct information in the transaction record; Col 13, Row 61 – Col 14, Row 3, by associating existing accurate merchant listing with transaction, user / cardholder financial institution can access additional information associated with the merchant to more accurately and specifically perform tasks such as identifying and categorizing transactions associated with the merchant; per Col 11, Rows 38-42, the transaction includes consumer credit card information; i.e., historical electronic transaction records comprises transaction information such as client / consumer’s credit card information and merchant portion ).
Regarding Claims 5 and 15, Barbour as modified by Evans discloses wherein the instructions to receive the transaction rule from the particular client comprises instructions to:
present a graphical user interface for client input (Evans, ¶17 and ¶28 and Fig 2A, user device 102 with electronic user interface for interactive display supported by output circuit 212 and to receive user input from the user via input circuit 210);
allow the particular client to select from preconfigured rule templates (Evans, ¶38 provide predefined / automatically generated rules to the user to include user defined exceptions in which the user defines situations in which a rule may be ignored); and
dynamically update the rule based on client-specific requirements (Evans, ¶39, upon receiving user defined automatically generated rules, index the rules and store the rules in the rules database 252).
Regarding Claims 7 and 17, Barbour discloses wherein the instructions to input the noisy data field into the named-entity-identification model comprises instructions to:
tokenize the noisy data field into substrings (Col 9, Rows 35-40, determine merchant portion of the transaction record, tokenize the merchant portion to determine terms to use as a query; see e.g., Col 17, Rows 10-33, parse character strings of electronic transaction record to determine merchant portion of received transaction record as search terms);
embed substrings into a vector in a multi-dimensional vector space (Col 17, Rows 34-53, obtain an embedding / single vector for the search terms from neural language modeler 267; per Col 22, Rows 26-27, neural language model maps terms into a multi-dimensional vector space); and
normalize the vector (Col 17, Row 53 – Col 18, Row 11, generate the single vector representing the search terms by weighting each of the search terms with inverse document frequency associated with respective terms).
Regarding Claims 9 and 19, Barbour as modified by Evans discloses wherein the instructions to apply the transaction rule comprises instructions to:
cross-reference the identified named entity with a list stored in a database and trigger an alert responsive to the transaction violates a predefined threshold (Evans, ¶37, alert user of a recurring payment for a good or service (i.e., pending transaction involving an entity) that violate one or more of the rules; ¶38, for example, after cross referencing a current merchant of the pending transaction to user defined exceptions to a general rule (i.e. whitelists) for particular merchants, execute the general rule to notify the user when an amount of time of disuse of the good or service provided by the current merchant exceeds a particular threshold value because the current merchant is not one of the particular merchants on the whitelists).
Regarding Claim 10, Barbour discloses wherein the named-entity-identification model is periodically updated to include:
new training data generated from cleared transactions (Col 16, Rows 19-27, update or additionally train a neural language model based on specific corpus comprising set of historical electronic transaction records; per Col 16, Rows 65-66, electronic transaction records associated with completed transaction); and enhanced features derived from recent merchant identification patterns (Col 13, Rows 50-61 and Col 25, Rows 6-22, electronic transaction records corresponding to “Thundercld Sub” corrected to the more accurate and specific “Thundercloud Subs”).
Claims 3 and 13 are rejected under 35 USC 103 for being unpatentable over Barbour (US 11640417 B2 A1) in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in further view of Relangi et al. (US 2020/0372219 A1).
Regarding Claims 3 and 13, Barbour discloses wherein the instructions to generate the training data from the past transaction data comprises instructions to:
identify patterns in noisy data fields based on predefined rules (Col 24, Rows 30-33, determine composite vector for merchant portion of the transaction record; Col 24, Rows 58-62, compare composite vector for the merchant portion in the transaction data to composite vector for each of the scoped set of merchant records to determine similarity between them; Col 25, Rows 1-5, identify the merchant record that is most similar to the merchant identification in the transaction data as the merchant associated with the transaction);
assign labels to named entities using historical data (Col 25, Rows 6-22, use the identified merchant record most similar to “Thundercld Sub” to determine the actual name of the merchant “Thundercloud Subs” and categorize (i.e., label) the transaction with the actual name “Thundercloud Subs” along with categorization as food, fast food, sandwich shop; note that the aforementioned comparison to merchant record is within the context of search merchant records based on electronic transaction records corresponding to the set of historical electronic transaction records of specific training corpus 269 per Col 16, Rows 17-20).
Barbour does not disclose validate the labeled data through a feedback loop.
Relangi teaches validating labeled data through a feedback loop (Abstract, Fig. 1, steps 130 -> 140 -> “No” -> 145; ¶37 and ¶43).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to valid labeled data through a feedback loop in order to receive corrected labelled named entities from a reviewer (Relangi, ¶10).
Claims 4 and 14 are rejected under 35 USC 103 for being unpatentable over Barbour (US 11640417 B2 A1) in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in view of Zhong et al. (US 2021/0303638 A1) and Li et al. (US 2022/0083744 A1).
Regarding Claims 4 and 14, Barbour discloses wherein the instructions to train the named-entity-identification model comprises instructions to:
utilize machine learning techniques with labeled training samples (Col 16, Rows 17-27, update or additionally train a neural language model based on specific corpus comprising a set of historical electronic transaction records with merchant records; per Col 25, Rows 6-22, merchant records correctly categorized or labeled with accurate name of the merchant and specific categories).
Barbour does not teach utilize supervised learning techniques with labeled training samples; iteratively optimize the model using gradient descent; and evaluate model accuracy based on cross-entropy loss.
Zhong teaches training an embedding model by utilizing supervised learning techniques (¶2, using artificial neural networks, support vector machines, decision tress, naïve Bayes classifiers to analyze data sets2) with labeled training samples (¶58, train embedding model 208 based on input string-entity pairs and labels 214); optimize the model using gradient descent and evaluate model accuracy based on cross-entropy loss (¶58, use gradient descent and backpropagation and a cross entropy loss function to update parameter values of embedding model 208 to reduce error between output of the embedding model 208 and corresponding labels 214).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to supervised learning techniques with labeled training samples, optimize the model using gradient descent, and evaluate model accuracy based on cross-entropy loss in order to reduce error between output of the neural language model / embedding model and corresponding labels / categorizations in the training corpus (Zhong, ¶58).
The combination does not teach that the optimization of the model using gradient descent is iterative.
Li teaches iterative optimization through gradient descent training of a neural language model (¶24).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to perform iterative optimization of the model using gradient descent in order to iterate until there is no further improvement in the loss function (Li, ¶24).
Claims 6 and 16 are rejected under 35 USC 103 for being unpatentable over Barbour (US 11640417 B2 A1) in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in view of Yang et al. (US 2012/0278705 A1).
Regarding Claims 6 and 16, Barbour does not disclose wherein the instruction to identify the noisy data field comprises instructions to apply heuristic algorithms to detect irrelevant substrings and standardize a format of extracted data fields.
Yang discloses automatically extracting meta data from unstructured electronic documents (Abstract) by applying heuristic algorithms to detect irrelevant substrings corresponding to noisy data fields (¶23, noise text line may be a text line mistaken as text line associated with header of an unstructured electronic document; ¶57, using heuristic approach to identify noise text line) and standardize a format of extracted data fields (¶57, classify text lines into different categories such as names (i.e., named entities), affiliations etc.).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to applying heuristic algorithms to detect irrelevant substrings corresponding to noisy data fields and standardize a format of extracted data fields in order to convert unstructured text into plain text (Yang, ¶57).
Claims 8 and 18 are rejected under 35 USC 103 for being unpatentable over Barbour (US 11640417 B2 A1) in view of Jamkhedkar et al. (US 2022/0156750 A1) and Evans et al. (US 2020/0184434 A1) as applied to claims 1 and 11, in view of Wu et al. (US 2020/0082928 A1).
Regarding Claims 8 and 18, Barbour discloses wherein the named-entity-identification model utilizes a neural network architecture (Col 16, Rows 1-8, use neural language model 267 to capture semantic and syntactic characterization of noisy terms in electronic transaction records; Col 22, Rows 13-14, neural network is used to determine numerical vectors corresponding to retrieved records).
Barbour does not teach the neural network architecture comprising: convolutional layers for feature extraction; recurrent layers for temporal pattern recognition; and fully connected layers for classification.
Wu discloses that it is well known to use a neural network architecture comprising convolutional layers for feature extraction (¶135, CNN encoding layer 1246 to generate vector representation of image input such as image messages to generate corresponding text per ¶59), recurrent layers for temporal pattern recognition (¶¶126-27, RNN-GRU encoding layer 1244 to project temporal voice inputs into vector representation), and fully connected layers for classification (¶148) of user inputs into vector space (¶121).
It would’ve been obvious to one ordinarily skilled in the art before the effective filing date of the invention to implement a neural network architecture comprising convolutional layers for feature extraction; recurrent layers for temporal pattern recognition; and fully connected layers for classification of a message input comprising text, voice, and image into vector space (Wu, Fig. 12 and see ¶148). For example, such neural language model can be applied to transaction comprising text, voice, and image to generate respective multi-dimensional vectors.
Conclusion
Applicant's amendment necessitated the new grounds of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any extension fee pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to examiner Richard Z. Zhu whose telephone number is 571-270-1587 or examiner’s supervisor Hai Phan whose telephone number is 571-272-6338. Examiner Richard Zhu can normally be reached on M-Th, 0730:1700.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/RICHARD Z ZHU/Primary Examiner, Art Unit 2654 10/02/2026
1US 11423373 B1, Col 55, Rows 45-49, supervised learning algorithms include artificial neural networks, Bayesian statistics, support vector machines, decision trees, classifiers.
2US 11423373 B1, Col 55, Rows 45-49, supervised learning algorithms include artificial neural networks, Bayesian statistics, support vector machines, decision trees, classifiers.
