Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined
under the first inventor to file provisions of the AIA . This office action is in response
to applicant's communication of May 22, 2025. The rejections are stated below. Claims
1 and 6-8 are pending and have been examined.
Response to Amendment/Arguments
Applicant argues that the Specification provides sufficient disclosure for the claimed first variational autoencoder (VAE) and the credit score inference model, citing paragraphs [0034] through [0039] and [0044]. Examiner acknowledges that the specification mentions a VAE model, a prediction network, and a gradient boosting decision tree framework. However, the written description requirement demands that the specification describe the claimed invention in sufficient detail to demonstrate that the inventor possessed the full scope of the claimed subject matter at the time of filing. Ariad Pharms., Inc. v. Eli Lilly & Co., 598 F.3d 1336, 1351 (Fed. Cir. 2010).
The claims recite specific limitations: "determining an inferred attribute using a first variational autoencoder (VAE) operating on at least the factual attribute," and "inferring, by a credit score inference model, a credit score to be set to the user based on the factual attribute and the inferred attribute." The specification describes that the machine learning unit generates and updates a VAE model by unsupervised learning using training data including factual attribute data, and that the inferred attribute determining unit obtains output values from multiple VAE models. But the specification does not disclose the specific architecture of the VAE, such as the number of layers, the dimensionality of the latent space, the encoder and decoder structures, or the training procedure beyond a general statement of unsupervised learning. Likewise, for the credit score inference model, the specification mentions a gradient boosting decision tree (LightGBM) but does not disclose the number of trees, the learning rate, the maximum depth, the feature sampling, or any other hyperparameters that would enable a person of ordinary skill to practice the claimed model without undue experimentation. The specification provides only a high level description of these models, not the detailed disclosure that the written description requirement contemplates.
Applicant cites In re Wertheim for the proposition that the Examiner bears the initial burden of presenting evidence or reasoning. Examiner has provided reasoning: the absence of specific algorithmic and architectural details means that the Specification does not reasonably convey possession of the claimed invention. A person of ordinary skill in the art, upon reading the specification, would not know which VAE architecture or which credit score inference model configuration was intended, because the Specification does not define these structures with particularity. The mention of a VAE and a gradient boosting decision tree does not constitute a written description of the specific claimed components, especially when the claims are directed to these components as structural elements of the information processing apparatus.
Applicant also argues that the specification discloses the training paradigms and the input output relationships but the written description requirement is not satisfied by describing what the models do; the requirement is satisfied by describing how the models are constructed or configured to perform those functions. The specification does not provide such construction details for the first VAE or the credit score inference model. Therefore, the rejection under 35 U.S.C. § 112(a) is maintained.
applicant argues that the claims are directed to a practical application and a technological improvement, citing Enfish and the USPTO memorandum on AI and ML claims. The Office has reviewed these arguments but finds them unpersuasive.
The claims are directed to a method and apparatus for inferring a credit score based on factual and inferred attributes. This is an abstract idea because it involves collecting and processing information about a user to evaluate creditworthiness. The concept of determining a credit score based on attributes is a fundamental economic practice, similar to evaluating credit risk, which has been recognized as an abstract idea. The additional elements, namely the use of a VAE and a credit score inference model, are computational tools that perform their ordinary functions of processing input data and generating output scores. The claims do not recite a specific technical improvement to the functioning of a computer or to another technology. The Specification asserts that the system improves inference accuracy, but it does not demonstrate how the particular architecture achieves that improvement or provide data showing a measurable enhancement over prior systems. The improvement described is an improvement in the accuracy of a credit score, which is an abstract result, not an improvement in computer technology or a technical field.
Applicant argues that the coordinated architecture of multiple VAEs and inference models addresses a technological problem of fragmented factual attributes. However, the problem of using fragmented data to evaluate credit is a business or informational problem, not a technological one. The solution as claimed, is to use machine learning models to process data, which is a application of existing computational techniques to an abstract idea. The claim does not recite a specific, unconventional arrangement of the models or a particular data transformation that would constitute a technological improvement. The use of a VAE and a credit score inference model, without more, does not integrate the abstract idea into a practical application because these models are used in their ordinary manner. Examiner refers to the guidance in MPEP § 2106.04(d)(1) and the Federal Circuit's decisions in Alice and Mayo. The claims do not recite additional elements that amount to significantly more than the abstract idea itself.
Applicant also relies on the USPTO's December 2025 memorandum regarding Ex Parte Desjardins. That memorandum reminds examiners to consider Enfish and other precedents when evaluating AI and ML claims. In this case, even under that guidance, the claims do not demonstrate a technological improvement. Enfish involved a self referential data table that improved computer functionality; the claims here do not improve the operation of the computer or the machine learning models themselves; they simply apply existing models to a new type of data. The improvement, if any, is in the accuracy of credit scoring, which is a business outcome, not a technical one. Therefore, the rejection under 35 U.S.C. § 101 is maintained for claims 1, 6, and 8 as directed to an abstract idea without significantly more.
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 and 6-8 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea of information processing without significantly more.
The Examiner has identified independent method Claim 7 as the claim that represents the claimed invention for analysis.
Claim 7 is directed to a method which is one of the four statutory categories of invention (Step 1: NO).
Claim 7 recites “a method performed by a …, the method comprising: accepting, as an input, a factual attribute that can be confirmed to be a fact with respect to a user based on user-provided data having been provided by the user oneself or history data of the user; determining, using a … operating on at least the factual attribute, an inferred attribute with respect to the user; and inferring, by a credit score inference model, a credit score to be set to the user based on the factual attribute and the inferred attribute, wherein the inferring comprises:
inferring the credit score to be set to the user based on an output value obtained
by inputting the factual attribute and the inferred attribute, to the credit score inference
model,
wherein the credit score inference model has been generated and/or updated
using teacher data based on : i) the factual attribute and the inferred attribute being an
input value, and ii) the credit score determined based on a payment history of deferred
payment related to users who share the factual attribute and the inferred attribute being
an output value,
determining the inferred attribute based on an output value obtained by inputting
the user-related data including the factual attribute to the …,
determining that, when an output value obtained from the … within a prescribed
range, the user has the inferred attribute, wherein the … uses the user-related data as
input value and produces a value indicating a probability of the user having a prescribed
inferred attribute as the output value,
wherein the method further comprises implementing a plurality of …, the plurality
of … including the … wherein, when the user-related data including the factual attribute
are input to the plurality of …, a plurality of vector representations with respect to the
user are obtained as an output value,
wherein the method further comprises implementing a plurality of inference
models wherein, when the plurality of vector representations of the user obtained from
the plurality of … are concatenated and input to the plurality of inference models, a value indicating a probability of the user having the prescribed inferred attribute is
obtained as an output value,
wherein the plurality of … are … with unsupervised learning,
wherein the plurality of inference models are …with the teacher data, and
wherein the teacher data comprises inferred attribute”. These limitations describe
an abstract idea of information processing and corresponds to Certain Methods of Organizing Human Activity (managing person behavior or relationships and following
rules or instructions or commercial or legal interactions). Accordingly, claim 7 recites an
abstract idea (Step 2A: Prong 1: YES).
The claim also recites as additional elements such as “computer”, “a first
variational autoencoder (VAE)”, “implementing a plurality of VAEs”, and “trained” which
do no more than implement the abstract idea and/or provide a particular technological
environment. Therefore, claim 7 recites an abstract idea without a practical application
(Step 2A - Prong 2: NO).
Further, as the additional elements of claim 7 do no more than serve as a tool to
implement the abstract idea and/or provide a particular technological environment, they
do not improve computer functionality or improve another technology or technical field.
Thus, claim 7 is not patent eligible (Step 2B: NO).
Claims 1 and 8 also recite the abstract idea of information processing and
corresponds to Certain Methods of Organizing Human Activity (commercial interactions
or sales activities or behaviors, business relations, managing personal behavior or
relationships or interactions between people) step one of step 2A (MPEP 2106.04).
Claim 1 includes the additional elements of “an information processing apparatus,
comprising: at least one memory configured to store program code; at least one
processor configured to operate as instructed by the program code, the program code
including: input code which, when executed causes the at least one of the at least
processor ..“, “determining code which, when executed, causes the at least one of the at
least one processor …”, “a first variational autoencoder (VAE) …”, “inference code,
which when executed, causes the at least one of the at least one processor to …”, “…
inference code is further configured to cause the at least one the at least one processor
to implement a plurality of VAEs”, “wherein the inference code is further configured to
cause the least one of the at least one processor ...”, “… the plurality of VAEs are
trained …”. Claim 8 includes the additional elements of “non-transitory computer
readable recording medium having recorded thereon program code to be executed by
at least one processor, the program code including: input code which, when executed
causes the at least one of the at least processor ..“, “determining code which, when
executed, causes the at least one of the at least one processor …”, “a first variational
autoencoder (VAE) …”, “inference code, which when executed, causes the at least one
of the at least one processor to …”, “… inference code is further configured to cause the
at least one the at least one processor to implement a plurality of VAEs”, “wherein the
inference code is further configured to cause the least one of the at least one processor
...”, “… the plurality of VAEs are trained …. The additional elements do no more than serve as a tool to implement the abstract idea and/or provide a particular technological environment. There is no improvement to the functioning of a computer, or lo any other technology or technical field (MPEP 2106.05(a}.
Claim 6 recites “wherein the … which, when … causes the at least one of the at least …, to further perform … infers the credit score using the credit score inference model generated and/or updated using a machine learning framework based on a gradient boosting decision tree” which further describe the abstract idea. The claim includes “wherein the inference code which, when executed, causes the at least one of the at least one processor” as an additional element. The additional element does no more than serve as a tool to implement the abstract idea and/or provide a particular technological environment. There is no improvement to the functioning of a computer, or to any other technology or technical field (MPEP 2106.05(a).
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
Claims 1, 6, and 8 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for pre-AIA the inventor(s), at the time the application was filed, had possession of the claimed invention.
Claims 1 and 8 each recite "code," "determining code," and "inference code". Applicant’s specification does not disclose the specific algorithm, architecture, or training methodology for the "first variational autoencoder (VAE)" or the "credit score inference model" that perform the core functions of determining an inferred attribute and inferring a credit score. In other words, the algorithms or steps/procedures taken to perform the function must be described with sufficient details so that one of ordinary skill in the art would understand how the inventor intended the functions to be performed. The claim recites these components in functional terms, but the specification lacks a description of how they are constructed and operate to achieve these results. 11. Claim 6 is rejected as it depends on claim 1.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KEVIN T POE whose telephone number is (571)272-9789. The examiner can normally be reached on Monday-Friday 9:30am through 6pm EST.
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, Ryan Donlon can be reached on 571-270-3602. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
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 https://ppair-my.uspto.gov/pair/PrivatePair. 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.
/K.T.P/Examiner, Art Unit 3692 /KEVIN T POE/
/RYAN D DONLON/Supervisory Patent Examiner, Art Unit 3692
Prosecution Insights