Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Objections
Claim 25 objected to because of the following informalities: claim 25 depends on claim 24. However, claim 24 has been canceled and amended as part of claim 21. Appropriate correction is required. For examination purpose, the examiner will treat claim 25 as dependent on claim 21.
Response to Amendment
The amendments filed 01/22/2026 have been entered. Claims 21-23, 25-27 remain pending
in the application.
Applicant’s amendments and arguments, with respect to claim rejections of claims 21-28 under 35 U.S.C 103 filed 09/25/2025 have been considered and they are persuasive. Therefore, the previous rejections as set forth in the previous office action will be removed.
Applicant’s amendments and arguments, with respect to claim rejections of claims 1 under 35 U.S.C 101 filed 09/25/2025 have been considered and they are not persuasive. Therefore, the previous rejections as set forth in the previous office action will be maintained.
The applicant argues that the amended claim 21 is patent-eligible because it is directed not merely to an abstract mathematical concept, but to specific technological improvement in knowledge graph reasoning. In particular, Applicant contends that the claim recites a Bayesian few-shot learning framework that builds a Gaussian mixture model based on entities and relations in a knowledge graph to reduce uncertainty, performs meta-training for newly appearing entities, constructs a meta learner using a Bayesian graph neural network, and trains the meta learner using a support set and posterior-distribution-based weight sampling so that the system can infer knowledge for new entities without fine-tuning or retraining. Applicant further asserts that these claimed features improve scalability, account for uncertainty in dynamic knowledge graph, and enable adaptive reasoning for newly emerging entities, which Applicant characterizes as an improvement in computer functionality rather than mere abstract reasoning. Applicant also relies on the specification’s discussion of deficiencies in prior approaches and analogizes amended claim 21 to claim 2 of UPSO July 2024 Subject Matter eligibility example 48, arguing that the claim therefore recites significantly more than a judicial exception.
The examiner respectfully disagrees. Applicant’s argument is not persuasive because amended claim 21, when considered as a whole, remains directed to a judicial exception, namely mathematical concepts and abstract analytical processing of knowledge graph data, and does not integrate that exception into a practical application. In particular, the claim recites representing entities and relations using probability distribution, representing conversion between entities using a Gaussian expression, constructing a Bayesian graph neural network using mathematical formulations, computing a gradient in a loss function, sampling a weight value from a posterior distribution, and training a meta learner based on those mathematical operations. Thus, much of the claim language defines the recited machine-learning model itself in terms of mathematical relationships and formulas. These limitations are directed to mathematical modeling and inferential reasoning over data rather than to a specific technological improvement in computer functionality.
Applicant’s asserted improvement is likewise not persuasive. Although Applicant contends that the claim improves knowledge graph reasoning, reduces uncertainty, improves scalability and enables inference for newly appearing entities without fine-tuning or retraining, the claimed output and alleged improvement are obtained from performing the recited mathematical calculations themselves. The alleged improvement resides in the abstract mathematical formulation of the machine-learning model and in the analytical result produced by that formulation, rather than in any recited improvement to computer architecture, processor operation, memory usage, network functionality, data storage, or other technological implementation. An improvement to the mathematical model or to the quality of the model’s inferential output without more, does not amount to significantly more than the abstract idea.
Further, the additional elements do not integrate the judicial exception into a practical application. The claim merely recites configuring a computer processor to perform the foregoing operations and applying the mathematical framework to entities and relations in a knowledge graph. The claim does not recite a particular machine implementation, specialized hardware arrangement, improved data structure, or other concrete technological mechanism that improves the functioning of the computer itself or another technology. Instead, the generic processor is used only as a tool to execute the recited mathematical calculations and model-training operations. Accordingly, the claim is directed to improving the abstract idea itself, not to a technological improvement.
Applicant’s reliance on the specification’s discussion of deficiencies in prior approaches and on USPTO July 2024 Subject Matter Eligibility Example 48 is also unpersuasive. Eligibility must be determined based on the language actually recited in the claim. Here, amended claim 32 does not recite a specific technological solution that effects an improvement in computer functionality, but instead recites mathematical relationships, model training operations, and inferential reasoning over knowledge graph data. Accordingly, amended claim 21 does not recite significantly more than the judicial exception. Therefore, the rejection of claim 21 under 35 U.S.C 101 is maintained as well as dependent claims 22-23, 25-27 because they do not recite additional elements sufficient to render the claims patent-eligible.
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 21-23, 25-27 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding claim 21,
Step 1:
Claim 21 recites a method, one of the four statutory categories of patentable subject matter.
Step 2A, Prong I:
Claim 21 further recites the limitations of:
“... perform task sampling”. This limitation recites an abstract idea of a mental process. The process of task sampling can be performed mentally within a person’s mind.
“... conducing random reasoning” This limitation recites an abstract idea of a mental process. The process of perform random reasoning can be performed mentally within a person’s mind.
“... wherein step (a) comprises: representing the head entity, the relation, and the tail entity as:
e
h
~
N
(
μ
h
,
∑
h
)
,
r
~
(
μ
r
,
∑
r
)
,
e
t
~
N
(
μ
t
,
∑
t
)
, respectively, where eh represents the head entity, r represents the relation, et represents the tail entity, μ represents location of the entity or the relation in a vector space, and ∑ represents a covariance whose magnitude is in positive correlation with uncertainty of the relation or the entity;” This limitation recites an abstract idea of a mathematical concept. The limitation recites a process of building a Gaussian mixture model based on mathematical equation that construct the model.
“representing conversion from the head entity to the tail entity as
(
e
t
-
e
h
)
~
N
(
μ
t
-
μ
h
,
∑
h
+
∑
t
)
;” This limitation recites an abstract idea of a mathematical concept. The limitation recites conversion function from entity to another entity using mathematical equation.
“wherein step (c) comprises: Constructing the meta learner based on the Bayesian neural network and relations in the knowledge graph, wherein the meta learner is represented as:
f
θ
=
1
n
e
i
(
S
i
)
∑
(
r
,
e
)
∈
n
e
i
(
S
i
)
B
[
r
|
|
e
]
, wherein,
f
(
θ
) represents a weight value, ß represents the Bayesian neural network, |
n
e
i
(
S
i
)
| represents a number of relation-entity pairs connected to the entity ei’, r represents the relation, e represents the tail entity;” This limitation recites an abstract idea of a mathematical concept. The limitation recites a neural network representing by mathematical equation.
“wherein steps (d) comprises: computing a gradient in a loss function based on the query set and a negative query set,” This limitation recites an abstract idea of a mathematical concept as well as a mental process. The limitation recites a mathematical concept of computing gradient in a loss function, where in a person can mentally perform the computing of value according to the concept.
“...
e
i
'
=
f
θ
~
P
r
θ
D
S
i
, wherein,
f
(
θ
) represents the weight value sampled from the posterior distribution, thus ensuring uncertainty and realizing random reasoning, Si represents the support set, and ei' represents a newly appearing entity” This limitation recites an abstract idea of a mathematical concept. The limitation recites a formula to calculate the weight value corresponding to an entity within the knowledge graph, which constitutes a mathematical formulation within the mathematical concept.
Step 2A, Prong II:
Claim 21 recites the following additional elements:
“configuring a computer processor to perform the step of” This additional element recites a generic computer element as a tool to perform the calculation with the abstract idea and does not significantly more than the abstract idea.
“(a) building a Gaussian mixture model based on entities and relations in a knowledge graph so as to reduce uncertainty of the knowledge graph”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The building of the Gaussian mixture model relies on the abstract idea of mathematical concept as recited in step 2A Prong I. The limitation does not demonstrate an improvement to a computer or a configuration of the model other than obtaining the model based on the mathematical calculation of the abstract idea and claiming the mathematical benefits but not a technological integration, thus make the claim does not significantly more than the abstract idea.
“(b) taking each said entities as a task to simulate a meta-training process of a newly appearing entity in the dynamic knowledge graph...” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f). The limitation merely recites the applying of knowledge graph to simulate the meta-training of entities, which represent a task. The limitation does not recite the configuration of the simulated training, or the implementation of the knowledge graph configuration to learn the entity but merely recite the usage of the knowledge graph, thus make the claim does not significantly more than the abstract idea.
“(c) constructing a meta learner based on a graph neural network and ...”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The limitation recites constructing the meta learner based on a graph neural network, wherein the meta learner configured based on the neural network is represented as the mathematical concept in step 2A Prong I. The limitation does not demonstrate an improvement to a computer function or a configuration of the neural network other than representing a neural network model through an abstract idea, thus make the claim does not significantly more than the abstract idea.
“(d) training the meta learner so as to use a support set to represent the newly appearing entity”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The limitation recites training the meta learner configured based on the neural network, wherein the neural network is represented as a mathematical concept in step 2A Prong I. The limitation does not recite how the training is performed or demonstrate an improvement to a machine learning algorithm. The above limitation recites a configuration of the meta learner based on the neural network being a mathematical concept such that the training the meta learner involves performing the calculation based of the mathematical concept, thus make the claim does not significantly more than the abstract idea.
“modeling based on the weight value f(θ), and by learning prior distribution following the Bayesian neural network, and reasoning uncertainty of newly emerging entities”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The limitation recites modeling based on the mathematical concept of f(θ), wherein f(θ) is demonstrated as a mathematical concept involve the meta learner as recited in step 2A Prong I. The claim simply recites the application of the mathematical concept as recited above without reciting how the modeling is configured or demonstrate an improvement to a computer, thus make the claim does not significantly more than the abstract idea.
“... updating and optimizing parameters of a reasoning model of the knowledge graph based on gradient descent data” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The limitation recites the application of the abstract idea of computing gradient data above to update and optimize the reasoning model without reciting how the update or optimization is configured or the specific technique to update and optimize the model, thus make the claim does not significantly more than the abstract idea.
“training the meta learner with the support set Si taken as an input and a representation of ei’ taken as an output” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The limitation recites the training of the meta learner, which is applied via the abstract idea of mathematical formulation above, which does not provide a specific machine learning technique or specific configuration to implement the training, but simply the application of the abstract idea, thus make the claim does not significantly more than the abstract idea
“upon completion of training with the support set Si, the meta learner uses the support set Si to represent real, new entities without fine-tuning and re-training.” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea. The limitation recites the application of the abstract idea of mathematical formulation above to represent the intended result, without providing specific implementation technique, or specific machine learning configuration step to implement the representation, while claiming the mathematical benefits but not a technological integration, thus make the claim does not significantly more than the abstract idea.
Step 2B:
When considered individually or in combination, the additional limitations and elements of claim 1 does not amount to significantly more than the judicial exception for the same reasons discussed above as to why the additional limitations do not integrate the abstract idea into a practical application. The additional elements of outlined in Step 2A performing functions as designed simply accomplishes execution of the abstract ideas.
“(a) building a Gaussian mixture model based on entities and relations in a knowledge graph so as to reduce uncertainty of the knowledge graph”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above
“(b) taking each said entity as a task to simulate a meta-training process of a newly appearing entity in the dynamic knowledge graph...” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above
“(c) constructing a meta learner based on a graph neural network and ...”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above
“(d) training the meta learner so as to use a support set to represent the newly appearing entity, ...”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above
“modeling based on the weight value f(θ), and by learning prior distribution following the Bayesian neural network, and reasoning uncertainty of newly emerging entities”. This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above
“... updating and optimizing parameters of a reasoning model of the knowledge graph based on gradient descent data” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above.
“training the meta learner with the support set Si taken as an input and a representation of ei’ taken as an output” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above.
“upon completion of training with the support set Si, the meta learner uses the support set Si to represent real, new entities without fine-tuning and re-training.” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not amount to significantly more than the judicial exception for the same reasons discussed above.
In conclusions from above for the elements considered as a mental process, elements reciting a mere instruction to apply an exception as identified in MPEP 2106.05(f) are carried over and do not provide significantly more than the abstract idea. Looking at the limitations in combination and the claims as a whole does not change this conclusion and the claim is ineligible.
Therefore, additional limitations of claim 21 do not amount to significantly more than the judicial exception.
Thus, claim 21 recites abstract ideas with additional elements rendered at a high level of generality resulting in claims that do not integrate the abstract idea into a practical application or amount to significantly more than the judicial exception.
Therefore, claim 21 is not patent eligible.
Regarding claim 22 depends on claim 21, thus the rejection of claim 21 is incorporated.
Claim 22 recites the limitation:
“The method for knowledge graph reasoning based on Bayesian few-shot learning according to Claim 21, wherein the score function used to compute the reliability of the triple is: S (eh, r, et) = KL (Pr, Pe), where, s represents the score function of the triple, eh represents the head entity, r represents the relation, et represents the tail entity, KL represents Kullback-Leibler divergence, Pr represents relation distribution, and Pe represents transformation distribution” This limitation recites an abstract idea of a mathematical concept. The limitation recites a process of compute the reliability using a score function, which involves mathematical equation. The calculation of such mathematical equation further involves abstract idea of a mental process.
Thus, claim 22 recites abstract ideas resulting in claims that do not integrate the abstract idea into a practical application or amount to significantly more than the judicial exception. Therefore, claim 22 is not patent eligible.
Regarding claim 23 depends on claim 22, thus the rejection of claim 22 is incorporated.
Claim 23 recites the limitation:
“wherein the processor is configurated for performing the step of taking each said entity as a task to simulate a meta-training process of a newly appearing entity in the dynamic knowledge graph and perform task sampling” This additional element recites an additional element of a mere instruction to apply an exception with a recitation of the words "apply it" (or an equivalent) as identified in MPEP 2106.05(f), and does not provide significantly more than the abstract idea or integration into a practical application. The limitation recites using a processor configurated to performing abstract idea of taking each said entity as a task to simulate a meta-training process without reciting the technical steps to perform the function by the processor. The limitation simply recites a high-level recitation of generic computer component used as a tool, and does not provide integration into a practical application or significantly more than the abstract idea.
“partitioning an original dataset into at least a meta training dataset that contains simulated newly appearing entity and a meta testing dataset that contains actual newly appearing entity” The limitation recites an abstract idea of a mental process. The process of partitioning a dataset into two datasets can be performed mentally within a human mind or manually using pen and paper.
“sampling the simulated newly appearing entity based on a meta-training process of the meta training dataset” The limitation recites an abstract idea of a mental process. The process of sampling an entity based on the training process of a training data set can be performed mentally within a human mind or manually using a pen and paper.
“maximizing a score of the triple of the query set based on a score function” The limitation recites an abstract idea of a mental process. The process of using a score function to calculate a maximizing score can be performed mentally within a human mind.
Thus, claim 23 recites abstract ideas with additional elements rendered at a high level of generality resulting in claims that do not integrate the abstract idea into a practical application or amount to significantly more than the judicial exception. Therefore, claim 23 is not patent eligible.
Regarding claim 25 depends on claim 21, thus the rejection of claim 21 is incorporated.
Claim 25 recites the limitation:
“minimizing KL divergence between the prior distribution and posterior distribution:
L
θ
*
=
min
θ
K
L
(
q
θ
|
|
Pr
θ
D
)
)
;
” The limitation recites an abstract idea of a mathematical concept. The limitation recites a mathematical function to minimize KL divergence.
“so that an objective function is represented as:
L
θ
*
=
K
L
(
q
θ
|
|
Pr
(
θ
)
)
-
E
θ
~
q
θ
[
log
P
r
(
D
|
θ
)
]
;” The limitation recites an abstract idea of a mathematical concept. The limitation recites a mathematical function of an objective function.
“wherein,
L
θ
represents the objective function,
q
(
θ
)
represents a hypothesis distribution for fitting
Pr
(
θ
)
,
Pr
(
θ
)
represents real distribution of the parameter,
Pr
θ
|
D
represents the posterior distribution and
Pr
D
|
θ
represents distribution of the training dataset.” The limitation recites an abstract idea of a mathematical concept. The limitation recites annotation for the mathematical function above.
Thus, claim 25 recites abstract ideas resulting in claims that do not integrate the abstract idea into a practical application or amount to significantly more than the judicial exception. Therefore, claim 25 is not patent eligible.
Regarding claim 26 depends on claim 25, thus the rejection of claim 25 is incorporated.
Claim 26 recites the limitation:
“wherein the newly appearing entities in the meta training dataset Mtr include: the corresponding support set
S
i
=
{
(
e
i
'
,
r
j
,
e
j
)
}
j
=
1
N
and query set
Q
i
=
{
(
e
i
'
,
r
j
,
e
j
)
}
j
=
N
+
1
n
(
e
i
)
; ” The limitation recites an abstract idea of a mathematical concept. The limitation recites support set and query set of entities being represented by mathematical function.
“where, n(ei’) represents a number of triples adjacent to the newly appearing entity, N represents a few-shot size, and ei’ represents the newly appearing entity” The limitation recites an abstract idea of a mathematical concept. The limitation recites annotation for the mathematical function above.
Thus, claim 26 recites abstract ideas resulting in claims that do not integrate the abstract idea into a practical application or amount to significantly more than the judicial exception. Therefore, claim 26 is not patent eligible.
Regarding claim 27 which recites a system, one of the four statutory categories of patentable subject matter. The claim is rejected under the same rationale of claim 21, because the claim recites similar limitation and processing steps.
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 DUY TU DIEP whose telephone number is (703)756-1738. The examiner can normally be reached M-F 8-4:30.
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, Alexey Shmatov can be reached at (571) 270-3428. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DUY T DIEP/Examiner, Art Unit 2123
/ALEXEY SHMATOV/Supervisory Patent Examiner, Art Unit 2123
Prosecution Insights